Post on 08-Feb-2019
The effects of experimental forestry treatments on site
conditions: short response study from an oak-hornbeam
forest
Bence Kovács Corresp., 1, 2, 3 , Flóra Tinya 1 , Erika Guba 1 , Csaba Németh 3 , Vivien Sass 4 , András Bidló 4 , PéterÓdor 1, 3
1 Institute of Ecology and Botany, MTA Centre for Ecological Research, Vácrátót, Hungary2 Department of Plant Systematics, Ecology and Theoretical Biology, Eötvös Loránd University, Budapest, Hungary3 GINOP Sustainable Ecosystems Research Group, MTA Centre for Ecological Research, Tihany, Hungary4 Institute of Environmental and Earth Sciences, University of Sopron, Sopron, Hungary
Corresponding Author: Bence Kovács
Email address: kovacs.bence@okologia.mta.hu
Background
Forest management alters the forest site, however, information is still limited about how different
silvicultural treatments modify these conditions. In the past decades, besides rotation forestry, new
silvicultural systems were introduced, fulfilling the requirements of multipurpose forestry. In this study
we investigated the short-term effects of different forestry treatments on microclimate, litter and soil
conditions in a European oak-dominated forest.
Methods
A forest ecological experiment was established in a homogenous, managed, 80 years old, Quercus
petraea and Carpinus betulus dominated forest, in 2014. Five treatments of three different forestry
systems were installed following a complete block design in six replicates: clear-cutting with a circular
retention tree group as typical elements of the clear-cutting system, preparation cutting (partial harvest)
belonging to the shelterwood system, gap-cutting as a common tool of continuous cover forestry in
Europe and uncut control. Microclimate, litter and soil variables were measured systematically since
2014. Here we present the results of the analyses of the first growing season following the interventions
(2015).
Results
We found that there is strong treatment effect in the case of microclimate and litter varibles, but for soil
characteristics the impacts will presumably appear in longer term. The increment of total and diffuse
light was the greatest in clear-cutting, in gap-cutting the illuminance was intermediate, while light-levels
were lower and less variant in preparation cutting and retention tree group. Air and soil temperature as
well as vapor pressure deficit increased the most in clear-cutting; both means and variances were the
highest in this treamtment. Retention tree group could not buffer the means of the temperature
variables, but a small group of tree individuals was able to ameliorate the extremes of the microclimate.
Significant increase of soil moisture was measured as a consequence of gap-cutting and less
pronouncedly in clear-cutting. Similarly, litter pH and moisture were the highest in these treatment types.
Significant increment in soil pH was detected in retention tree group. Through the analysis of
microclimate variables during the growing season, we could demonstrate the buffering effect of forest
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
canopy: differences between treatments were the greatest in summer for all microclimate variables.
Discussion
We can conclude that in oak–hornbeam forest, only less intensive and spatially heterogeneous
silvicultural treatments could preserve the stable, cooler and humid below-canopy microclimate,
therefore, group selection using gaps and irregular shelterwood systems are favourable. Our findings can
support the mitigation of the negative impacts of climate change in managed forest. Moreover, besides
basic research we can formulate implications for foresters and conservationists to preserve biodiversity
in temperate forests.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
1 The effects of experimental forestry treatments on site conditions: short response study
2 from an oak-hornbeam forest
3
4 Kovács, B.1,2,3, Tinya, F.1, Guba, E.1, Németh, Cs.3, Sass, V.4, Bidló, A.4, Ódor, P.1,3
5
61 MTA Centre for Ecological Research, Institute of Ecology and Botany, Alkotmány út 2-4, H-
7 2163 Vácrátót, Hungary
82 Department of Plant Systematics, Ecology and Theoretical Biology, Eötvös Loránd University,
9 Pázmány Péter sétány 1/C, H-1117 Budapest, Hungary
103 MTA Centre for Ecological Research, GINOP Sustainable Ecosystems Research Group,
11 Klebelsberg Kuno utca 3, H-8237 Tihany, Hungary
124 University of Sopron, Institute of Environmental and Earth Sciences, Bajcsy-Zsilinszky utca 4,
13 H-9400 Sopron, Hungary
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
14 ABSTRACT
15 Background
16 Forest management alters the forest site, however, information is still limited about how different
17 silvicultural treatments modify these conditions. In the past decades, besides rotation forestry, new
18 silvicultural systems were introduced, fulfilling the requirements of multipurpose forestry. In this study
19 we investigated the short-term effects of different forestry treatments on microclimate, litter and soil
20 conditions in a European oak-dominated forest.
21 Methods
22 A forest ecological experiment was established in a homogenous, managed, 80 years old, Quercus
23 petraea and Carpinus betulus dominated forest, in 2014. Five treatments of three different forestry
24 systems were installed following a complete block design in six replicates: clear-cutting with a circular
25 retention tree group as typical elements of the clear-cutting system, preparation cutting (partial harvest)
26 belonging to the shelterwood system, gap-cutting as a common tool of continuous cover forestry in
27 Europe and uncut control. Microclimate, litter and soil variables were measured systematically since
28 2014. Here we present the results of the analyses of the first growing season following the interventions
29 (2015).
30 Results
31 We found that there is strong treatment effect in the case of microclimate and litter varibles, but for soil
32 characteristics the impacts will presumably appear in longer term. The increment of total and diffuse light
33 was the greatest in clear-cutting, in gap-cutting the illuminance was intermediate, while light-levels were
34 lower and less variant in preparation cutting and retention tree group. Air and soil temperature as well as
35 vapor pressure deficit increased the most in clear-cutting; both means and variances were the highest in
36 this treamtment. Retention tree group could not buffer the means of the temperature variables, but a small
37 group of tree individuals was able to ameliorate the extremes of the microclimate. Significant increase of
38 soil moisture was measured as a consequence of gap-cutting and less pronouncedly in clear-cutting.
39 Similarly, litter pH and moisture were the highest in these treatment types. Significant increment in soil
40 pH was detected in retention tree group. Through the analysis of microclimate variables during the
41 growing season, we could demonstrate the buffering effect of forest canopy: differences between
42 treatments were the greatest in summer for all microclimate variables.
43 Discussion
44 We can conclude that in oak–hornbeam forest, only less intensive and spatially heterogeneous
45 silvicultural treatments could preserve the stable, cooler and humid below-canopy microclimate,
46 therefore, group selection using gaps and irregular shelterwood systems are favourable. Our findings can
47 support the mitigation of the negative impacts of climate change in managed forest. Moreover, besides
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
48 basic research we can formulate implications for foresters and conservationists to preserve biodiversity in
49 temperate forests.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
50 INTRODUCTION
51
52 Forest management induces substantial alterations in environmental conditions that fundamentally
53 influence ecosystem structure and functions (Edwards et al., 2014; Zhao and Jackson, 2014).
54 Furthermore, these impacts also affect the long-term survival, regeneration and diversity of forest-
55 dwelling organism groups (e.g. Paillet et al., 2010; Chaudhary et al., 2016).
56 Silvicultural treatments cause changes in biodiversity both directly and indirectly (Keenan &
57 Kimmins, 1993; Rosenvald & Lõhmus, 2008). Direct effects are pronounced by the elimination of
58 food resources and indispensable microhabitats, such as (host) tree individuals that are substrate
59 of epiphytes (Lõhmus & Lõhmus, 2010), woody debris for deadwood-dependent communities
60 (Seibold et al., 2015) or standing dead trees as commodity for cavity nesters (Ibarra et al., 2017).
61 However, most species of the forest biota are influenced by forest management through indirect
62 pathways: through the alteration of forest site conditions – microclimate, litter attributes, soil
63 characteristics – and biogeochemical cycles (Zheng et al., 2000; Sayer, 2005; Thiffault et al., 2011;
64 Kishchuk et al., 2014; Frey et al., 2016). Microclimate is also a major driver of ecosystem
65 processes such as decomposition, respiration and nutrient dynamics (Thibodeau et al., 2000;
66 Stoffel et al., 2010; Knapp et al., 2014).
67 Studying the effects of different management types on forest site conditions – and especially on
68 microclimate – at local scale could provide evidences that support the adaptation strategies
69 mitigating the negative impacts of climate change (Suggitt et al., 2011; Latimer & Zuckerberg,
70 2017). Fine-scale measurements and models are necessary to calculate probable species
71 distributions and creating predictions for the ecological processes (De Frenne et al., 2013; Lenoir,
72 Hattab & Pierre, 2017; Greiser et al., 2018). Therefore, for conservation purposes, it is important
73 to investigate how forestry treatments alter the forest site conditions, through numerous, highly
74 interrelated effects.
75 Forest stands create unique, buffered below-canopy microclimates (Geiger, Aron & Todhunter,
76 1995; Chen et al., 1999), compared to open-fields (von Arx, Dobbertin & Rebetez, 2012) or to
77 plantations (Hardwick et al., 2015). It is mainly determined by the tree species composition, tree
78 species richness and the stand structure (Lin et al., 2017; Gebauer, Horna & Leuschner, 2012; von
79 Arx, Dobbertin & Rebetez, 2012; Ehbrecht et al., 2017): foliage of the different vegetation layers
80 absorbs a great proportion of incoming energy and reduces the loss of longwave radiation from
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
81 the surface. Soil and litter characteristics are also influential (von Arx et al., 2013; Kovács, Tinya
82 & Ódor, 2017). As a result of these effects forest microclimate can be characterized by lower
83 diurnal and seasonal magnitudes in case of several microclimatic variables (Geiger, Aron &
84 Todhunter, 1995; von Arx et al., 2013). Similarly, litter layer is important for many decomposer
85 groups as habitat and food resource, for the nutrient cycling, smoothing the fluctuations of the
86 physical and chemical properties of the topsoil and below-canopy ambient air (Ogée & Brunet,
87 2002; Sayer, 2005).
88 The alterations of the microclimate, litter and soil properties and thus changes in forest
89 communities are highly dependent on the spatial extent, the spatiotemporal pattern, frequency or
90 severity of the applied forest management approaches (Bicknell et al., 2014; Dieler et al., 2017;
91 Schall et al., 2018). Changes in the main structural elements of forests (e.g. canopy closure,
92 horizontal and vertical foliage distribution) result in considerable alterations in the processes of
93 the soil-vegetation-atmosphere system. This effect is unambiguous at clear-cuttings (Keenan &
94 Kimmins, 1993; Aussenac, 2000; Marshall, 2000), but it is also observable at partial cuttings like
95 thinning, group selection or retention tree harvesting (Carlson & Groot, 1997; Heithecker &
96 Halpern, 2006; Ryu et al., 2009; Bigelow & North, 2012; Kishchuk et al., 2014; Coulombe, Sirois
97 & Paré, 2017).
98 As Shall et al. (2018) stated, forest management in the temperate zone is shifting globally from
99 even-aged rotation management systems towards continuous cover silvicultural approaches that
100 support structural heterogeneity. Bernes et al.(2015) pointed out that there is a knowledge gap
101 concerning Central European deciduous forests in the aspect of possible multipurpose forest
102 management alternatives. Forestry experiments are necessary for the understanding of the complex
103 relationships between different management practices and forest site, regeneration and biodiversity
104 characteristics. There are several studies that examine the effects of multiple treatment levels
105 within a particular silvicultural system on microclimate: different artificial gap sizes and shapes
106 (Carlson & Groot, 1997; Gálhidy et al., 2006), various combinations of retention levels and spatial
107 configurations (Heithecker & Halpern, 2006), thinning intensities and patterns (Weng et al., 2007;
108 Rambo & North, 2009) and fuel-reduction oriented treatments (Ma et al., 2010). Similarly, several
109 research presented results of changes in soil characteristics and biochemical processes by the
110 investigation of lower intensity harvest types (Thibodeau et al., 2000; Ryu et al., 2009; Jerabkova
111 et al., 2011; Coulombe, Sirois & Paré, 2017)). Most of the studies dealing with the interactions of
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
112 harvesting and forest site conditions concentrate on one selected forestry system, whereas it is rare
113 to apply two or more silvicultural strategies within one experimental design (but see e.g. Kishchuk
114 et al., 2014; Knapp et al., 2014). The “Pilis Experiment” was implemented to compare the effects
115 of different treatment types – belonging to three opposing silvicultural systems eligible in Europe
116 (clear-cutting, shelterwood system and continuous-cover forestry with gap-cutting; Matthews,
117 1991) – on site conditions, regeneration and biodiversity. Our open-field, multi-taxa experiment
118 was established in 2014 (http://piliskiserlet.okologia.mta.hu/en). As a target habitat type, sessile
119 oak-hornbeam forests were chosen, which represent a widespread deciduous woodland habitat
120 type in the Pannon Ecoregion (Bölöni et al., 2008) and generally in Central Europe (Brus et al.,
121 2012). Furthermore, in line with that, this is one of the focal indigenous forest types for timber
122 harvesting in this region, due to the high-quality timber of sessile oak (Annighöfer et al., 2015).
123 Here, we focus on the effects of the performed experimental treatments on forest site conditions
124 during the first post-harvest period. Our objectives are to quantify differences induced by the
125 applied management treatments (clear-cutting, gap-cutting, preparation cutting and retention tree
126 group) (1) on the mean and magnitude of microclimate, soil and litter variables during the growing
127 season, (2) on the temporal pattern of site condition variables through a growing season, and (3)
128 on the diurnal patterns of selected microclimate variables.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
129 MATERIALS AND METHODS
130
131 Study area
132 The study was conducted in the Pilis Mountains, north-eastern ridge of the Transdanubian Range,
133 Hungary (47°40’N, 18°54’E; Fig. 1A). Plots are situated on a horst ridge (370–470 m a.s.l) on
134 moderate (7.0–10.6°), north-facing slopes. Average annual mean temperature is 9.0-9.5°C (16.0–
135 17.0°C in the growing season) with a mean annual precipitation of 650 mm (Dövényi, 2010).
136 The bedrock consists of Dachsteinian limestone and Lattorfian sandstone with loess (Dövényi,
137 2010). According to the soil profiles established on the study site, depth of the soil is changing
138 along a slight topographic gradient (Supplemental Information 1): deep (250 cm) at the lower parts
139 and shallow (70 cm) nearer the ridge. In the lower parts, the soil type is brown forest soil with clay
140 illuviation (luvisol), while in the upper parts it is rendzic leptosol (Krasilnikov et al., 2009). Soils
141 are slightly acidic (pH of the 0-20 cm layer is 4.6±0.2). The physical and chemical characteristics
142 of the upper 50 cm of soil are similar in the whole area independently from soil depth
143 (Supplemental Information 1). The variety of soil types did not cause discernible variability in the
144 woody vegetation (Table 1).
145 The study site was established in an approximately 40 ha sized homogeneous block in a managed,
146 two-layered sessile oak-hornbeam forest stand (Natura 2000 code: 91G0; Council Directive
147 92/43/EEC). The study site is legally protected, the experiment was approved by the Pest Megyei
148 Kormányhivatal (Pest County Administration; permission number: KTF:30362-3/2014). The
149 stand is even-aged (80 years old) and has a relatively uniform structure (Table 1) and species
150 composition due to the applied shelterwood silvicultural system. Upper canopy layer (average
151 height: 21 m, mean DBH: 27.6 cm) is dominated by sessile oak (Quercus petraea Matt. (Liebl.)),
152 while the second most abundant tree species, hornbeam (Carpinus betulus L.) forms a subcanopy
153 layer with an average height of 11 m and mean DBH of 11.6 cm (Table 1 and Fig. S2.2). Other
154 woody species are rare, individuals of Fraxinus ornus L., Fagus sylvatica L., Quercus cerris L.
155 and Cerasus avium L. were recorded in the tree layer as admixing tree species. Before the
156 experimental treatments, mean basal area (BA) of the upper layer was 29.4(±4.3) m2ha-1 and
157 8.8(±2.6) m2ha-1 in the case of the secondary canopy layer, respectively (Table 1). Canopy closure
158 was rather homogenous, it varied between 81 and 94%. Shrub layer was scarce and mainly
159 consisted of the regeneration of hornbeam and Fraxinus ornus L. with lower cover of shrub species
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
160 e.g. Crataegus monogyna Jacq., Cornus mas L., Ligustrum vulgare L. and Euonymus verrucosus
161 Scop.. The understory layer was formed by general and mesic forest species, dominant species are
162 Carex pilosa Scop., Melica uniflora Retz., Cardamine bulbifera L., Galium odoratum (L.) Scop.
163 and Galium schultesii Vest. Before the treatments (in 2014), cover of herb layer was approximately
164 40% (unpublished data).
165
166 Study design
167 Five treatment types (Fig. S2.1) were implemented in a randomized complete block design in six
168 replicates (hereafter blocks, Fig. 1B and 1C):
169 1. Control (C): The original stand characteristics remained unaltered.
170 2. Clear-cutting (CC): Approximately 0.5 ha sized circular clear-cuts were formed surrounded
171 by closed-canopy stand. The area of the treatment was designated as the area surrounded by the
172 trunks of the peripheral dominant forest trees: the applied diameter was 80 m. Within CC, every
173 tree individuals (DBH ≥ 5 cm and/or height ≥ 2 m) were cut, that caused drastic changes in tree
174 structure (BA change: -39.6 m2ha-1) and canopy closure (change: -85.4%).
175 3. Gap-cutting (G): Circular artificial gaps were established in the closed stand by the
176 elimination of all tree individuals within a diameter of 20 m (~0.03 ha). Gap size was defined
177 as expanded gaps (Runkle, 1981), i.e. by measuring the base of surrounding canopy trees. The
178 chosen 1:1 gap diameter/intact canopy height ratio is widely used in Central Europe for
179 transition system applying gap-cuttings and it also fits well for the records of gap area in oak
180 forests (Wijdeven & van Hees, 2001; Dey, 2002).
181 4. Preparation cutting (P): Uniform partial cutting was applied within a circle with a diameter
182 of 80 m; 30% of the initial total basal area of the upper canopy layer was cut and the felled trees
183 were distributed evenly (BA change: -8.4 m2ha-1). Furthermore, the complete subcanopy- and
184 shrub layer were also removed.
185 5. Retention tree group (R): All tree and shrub individuals were retained within a 0.03 ha sized
186 circular plot (diameter=20 m) in the clear-felled area, that resulted a small patch of remained
187 stand with approximately 8-12 trees of the former upper layer.
188 Clear-felled areas in the hilly region of Hungary are less than 5 ha according to the operative law
189 (Act No. LXVI of 2017). The created clear-cuttings are substantially smaller than it is typical in
190 Hungary or in the temperate deciduous forests in Europe (3-10 ha, Standovár, 2006). Therefore,
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
191 changes in the site conditions resulted by this treatment may be less pronounced and drastic.
192 Neither the plot of clear-cuttings, nor the retention tree groups were placed in the center of the
193 clear-cuts (Fig. S2.2). These plots were shifted to the 1:3 intersections along the east-west
194 diameter, because we intended to minimize the bias caused by the shading of the remained trees
195 of retention tree groups in the clear-cutting plots.
196 All treatments were carried out in the winter of 2014-2015. In the center of the treatments, a
197 6 × 6 m fenced area (hereafter plot) was established to exclude the effects of the large-bodied game
198 species.
199
200 Data collection
201 The “Pilis Experiment” follows BACI (Before-After-Control-Impact, Stewart-Oaten, Murdoch &
202 Parker, 1986) design, therefore, all measurements are systematically repeated as well as in the pre-
203 treatment year (2014) and the following post-treatment year (2015). Microclimate variables (total
204 light, air temperature, relative humidity, soil temperature and soil moisture) were recorded in every
205 month of the growing season, litter and soil variables (litter mass, litter pH, litter moisture content,
206 soil pH, hygroscopicity, nutrient content) were measured in two sampling periods per year.
207 Systematic microclimate measurements were taken place in the center of each plot. Temporally
208 synchronized data collections were carried out using 4-channeled Onset ‘HOBO H021-002’ data
209 loggers (Onset Computer Corporation, Bourne, US-MA) mounted on wooden poles (Fig. S2.3). In
210 every month through the growing season (March-October), 72-hour logging periods were applied
211 with 10-min logging intervals. Photosynthetically active radiation (PAR, λ=400-700 nm;
212 μEm−2s−1) was measured at 150 cm above ground level using Onset ‘S-LIA-M003’ quantum
213 sensors. Air temperature (Tair; °C) and relative humidity (RH; %) data were collected at 130 cm
214 above ground level with Onset ‘S-THB-M002’ combined T/RH sensors housed in standard
215 radiation shields to avoid direct sunlight. Soil temperature (Tsoil; °C) was measured with ‘S-TMB-
216 M002’ 12-Bit temperature sensors by Onset placed 2 cm below ground. Soil water content
217 (volumetric water content, SWC; m3/m3) data were collected by Onset ‘S-SMD-M005’ soil
218 moisture sensors buried 20 cm below ground level to measure the average soil moisture in 10-20
219 cm soil depth. Air temperature and relative humidity data were used to calculate vapor pressure
220 deficit (VPD; kPa) values at every logged occasion following the protocol of Allen et al. (1998):
221 VPD=(0.6108){exp[17.27·T/(237.3+T)]}·(1-RH/100). The reason of using VPD as a background
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
222 variable is that it can give a direct indication of the atmospheric moisture conditions independently
223 of the actual temperature. Therefore, it is a good standalone indicator of the atmospheric factors
224 influencing evaporation: VPD describes the actual drying capacity of the air, i.e. the higher the
225 VPD is, the more intensive is the evaporation (Anderson, 1936). Additionally, relative diffuse light
226 (DIFN; %) was measured by LAI-2000 Plant Canopy Analyzer (LI-COR Inc., Lincoln, US-NE)
227 in the center of each plot at 130 cm above ground level. Measurements were carried out in August
228 at dusk to avoid direct light getting into the sensor. Repeated measurements are not needed with
229 this device (Tinya et al., 2009). A 270° view restrictor masked the portion of the sky containing
230 the sun and the operator (LI-COR Inc., 1992). Reference above-canopy measuring was performed
231 on an adjacent open field.
232 At each plot, four litter (30 × 30 cm area) and topsoil (0-20 cm depth) samples were systematically
233 collected within the adjoining 3-meter sphere of the plots. The samples were taken twice a year:
234 in April and in October. All samples were returned to the laboratory and following the necessary
235 preparation steps, litter mass, litter pH, litter moisture content, Kuron’s hygroscopicity (hy), soil
236 organic matter content and nutrient content were measured. Litter mass was measured after air-
237 drying for 48 h. Litter moisture content (%) was calculated as the mass loss of the freshly collected
238 litter samples (i.e. the difference of the fresh and dried litter).
239 Litter and soil pH was potentiometrically measured using supernatant suspension of air-dried and
240 sieved (<2 mm) samples and 25 ml of distilled water, the applied mass was 5 g for litter and 10 g
241 for soil samples, respectively (MSZ-08-0205:1978). Kuron’s method was applied for gauging hy
242 of air-dry soils (Verstraeten & Livens, 1971): with 50% (v/v) H2SO4 solution and 35.2% RH
243 according to MSZ-08-0205:1978. Chemical compounds were evaluated on composite samples of
244 the 1:1 mixture of the four, sieved (<0.5 mm) subsamples per plot. Total soil carbon and nitrogen
245 content were determined by dry combustion method using Elementar vario MAX CNS-analyzer
246 (Elementar Analysesysteme, Langenselbold, Germany) applying the ISO standards (ISO
247 10694:1995; ISO 13878:1998): soil samples were weighed up to 80-100 g, and a tungsten oxide
248 catalyst was added. The applied combustion temperature was 1140°C. Plant available phosphorus
249 and potassium were determined by ammonium lactate (AL) solution method (0.1M NH4-
250 lactate + 0.4 M HOAc, adjusted to pH 3.75) developed by Egnér et al. (1960 cf. Carter &
251 Gregorich, 2008) according to the operative Hungarian standards (MSZ 20135:1999). PAL was
252 measured colorimetrically, KAL was quantified by flame photometry.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
253
254 Data analysis
255 Microclimate data were initially screened and obvious errors caused by technical failures
256 (indicated by e.g. unrealistic data or large spikes in variables), were replaced by missing values.
257 The manually corrected data were imported into the database built in SpatiaLite 4.3.0a (Furieri,
258 2015).
259 Firstly, observations were split into 24-hour datasets. Based on these diurnal data, on the one hand
260 descriptive statistics, as daily mean and daily interquartile range (IQR), were calculated for each
261 plot, on the other hand using the raw data, differences from the values collected at the control were
262 calculated for every recording (control values were subtracted from the treatment values of the
263 block) and then mean and IQR were computed. To measure the direct effects of the silvicultural
264 treatments on site condition variables, these relative data were used to avoid the effects of the
265 actual synoptic situation (in case of the microclimate variables) and to minimize spatio-temporal
266 heterogeneity of soil and litter variables. For the analyses, one randomly chosen 24-hr
267 microclimate dataset was used in every month. Daily IQR of SWC was excluded from the analysis,
268 because soil moisture is a rather stationary variable. For RH and VPD by virtue of numerous
269 missing data, the subset of October was excluded.
270 The temporal patterns of the measured variables were investigated using two different temporal
271 resolutions: according to the distinct methods for soil chemical variables and litter parameters
272 seasons were compared (spring vs. autumn), while in the case of microclimate variables we used
273 months as factor levels.
274 As a result of the relatively short measurement campaigns (3 days per month), the classical BACI
275 design was not suitable for data analysis, due to the possible different weather conditions.
276 Therefore, only the analyses based on the dataset of year 2015 are presented here, while results of
277 the pre-treatment datasets (2014) can be found in Supplemental Information 3.
278 To explore the effect of treatments and time on the measured site condition variables, linear mixed
279 models were used (Faraway, 2006). Data were transformed where it was necessary to achieve the
280 normality of model residuals. The effects of different treatment levels across (1) the whole growing
281 season and (2) over the applied temporal resolution (month or season) were tested by the same
282 modeling framework: treatment, time and their interaction were used as fixed factors, while block
283 as a random factor. Models’ goodness-of-fit values were measured by likelihood-ratio test-based
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
284 coefficient of determination (R2LR; Bartoń, 2016). Differences between treatment levels were
285 evaluated by all-pairwise comparisons of Tukey procedure (alpha=0.05): multiple comparisons
286 were applied for the differences among treatments through the growing season by general linear
287 hypotheses (Hothorn, Bretz & Westfall, 2008), while comparisons across treatments within the
288 time levels were performed among least-squares means (Lenth, 2016). The significance of
289 differences between control and the other treatment levels was tested by random effect models
290 without intercept (Zuur, 2009). The diurnal pattern was only analyzed qualitatively without any
291 statistical test applying standard LOWESS analyses with 95% confidence intervals. Datasets (raw
292 data) were pooled into two groups: the peak of the growing season (i.e. June, July, August) and
293 the transitional period (March, April, September, October) with lower canopy closure. Smoothing
294 procedures were applied on three or four 24-hr datasets with six replications of each treatment
295 levels.
296 Data analyses were performed with R version 3.4.1. (R Core Team, 2017). Mixed models were
297 conducted by R package ‘nlme’ (Pinheiro et al., 2017), multiple comparisons were appraised by
298 ‘multcomp’ (Hothorn, Bretz & Westfall, 2008) and ‘lsmeans’ (Lenth, 2016) packages,
299 determination coefficients of the mixed models were calculated by ‘rsquaredLR’ function of
300 ‘MuMIn’ package (Bartoń, 2016). For graphing means and SDs the modified script of ‘errorbars’
301 function was used (Reiczigel, Harnos & Solymosi, 2014).
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
302 RESULTS
303
304 The effects of experimental treatments on site condition variables
305 According to the performed linear mixed effect models, we found that experimental treatments
306 affect microclimate and litter variables more, while soil chemical characteristics did not differ
307 significantly among treatment types – except for topsoil acidity (Table 2, Fig. 2).
308 Both mean and variability of total as well as mean of relative diffuse light were substantially higher
309 in all treatments than in the controls (Fig. 2a, Fig. 2b and Fig. 2c, respectively). The largest values
310 were detected in clear-cutting and the increment was also considerable in gap-cutting. Preparation
311 cutting and retention tree group had similar light conditions, but diffuse light was lower in retention
312 tree groups. In retention tree groups and clear-cutting the mean of dTair were significantly higher
313 than in the other two treatments (Fig. 2d). Air temperature was buffered the most in the preparation
314 cutting, but the means of dTair were not different between gap-cutting and preparation cutting. The
315 interquartile range of air temperature also departed significantly from the control in all treatments
316 (Fig. 2e). In the clear-cutting, both the mean and the standard deviation of the IQR were the
317 highest. The lowest range was measured in the gap-cutting, while in the other treatments, IQR was
318 intermediate. dRHmean was the lowest in clear-cutting and retention tree group (Fig. 2f). In
319 preparation cutting and gap-cutting, humidity remained similar to control; furthermore, these
320 treatments did not differ from each other. It is noticeable that the range of dRH was the lowest in
321 gap-cutting, and highest in clear-cutting, however in all treatment, IQRs of dRH were departed
322 from control (Fig. 2g). The pattern of dVPD (both mean and IQR, Fig. 2h and Fig. 2i, respectively)
323 across treatment levels was similar to dTair (because of the high contribution of temperature to this
324 variable), and all treatments differed also significantly from control. Soil temperature, means just
325 as IQR differed significantly in every treatment from control (Fig. 2j, Fig. 2k, respectively). Clear-
326 cutting created soil thermal conditions that divaricated the most from the closed stand: both mean
327 and IQR of the dTsoil were the highest there. In retention tree group mean of dTsoil did not differ
328 significantly from clear-cutting, but IQR were significantly lower. The coolest soil environment
329 with the less diurnal variability was created by gap-cutting. Soil moisture differed significantly in
330 clear-cutting and even more in gap-cutting from the controls. The highest SWC was measured in
331 gap-cutting, the difference was smaller in the clear-cutting, while in preparation cutting and
332 retention tree groups a slight decrease was detected, and the latter was the driest treatment (Fig. 2l).
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
333 Litter variables showed almost as strong relationship with treatment levels as microclimate
334 variables. In the first year there was no significant response in litter mass, however, it showed an
335 increasing trend from clear-cutting to retention tree group (Fig. 2m). In clear-cutting and gap-
336 cutting litter pH departed significantly from that in control, and litter were more neutral in these
337 two treatments than in preparation cutting and retention tree group (Fig. 2n). Litter moisture
338 followed the sequence of dSWC: forest floor was the driest in retention tree group, but it was not
339 significantly different from that in control; and litter moisture was significantly higher in the other
340 three treatments – the highest values were measured in gap-cutting (Fig. 2o). Topsoil was less
341 acidic in preparation cutting than in the other treatments and it was the only treatment level where
342 soil pH differed from control (Fig. 2p).
343
344 Temporal differences among treatments through the growing season
345 Besides the study of the treatment effects, the temporal differences during a growing season were
346 also analyzed (Table 2). Except for light and soil moisture variables, the effect of time was similar
347 or stronger than that of treatments. For litter mass and potassium content, only the time effect was
348 significant.
349 The largest differences in microclimate variables were detectable in summer (Fig. 3). dPAR was
350 the highest in clear-cutting almost in every month, but the differences were highest in full-leaved
351 months – from May to August (Fig. 3a). dVPD values were highly divergent among treatment
352 levels during summer, the drying capacity of air was significantly higher in clear-cutting and in
353 the retention tree group, while dVPD did not depart substantially from control in the two other
354 treatment types (Fig. 3b). In the case of soil temperature (Fig. 3c), some important results could
355 be asserted: thermal input is the largest in clear-cutting with the highest variance during summer;
356 the differences between treatment levels accelerated as the canopy closure increased, but gap-
357 cutting and preparation cutting remained similar to the control during the whole growing season;
358 retention tree groups could partly buffer the heating through the shading of the remained trees; and
359 as the first frosts appeared, dTsoil differed greatly from the other treatments (i.e. soil was the coldest
360 in clear-cutting in October). The soil moisture increment in gap-cutting was detectable through the
361 whole measurement period, and it enhanced during summer. dSWC was also relatively high in
362 clear-cutting, but the difference was less pronounced than in gap-cutting (Fig. 3d). A moderate soil
363 desiccation was present in retention tree group from June to September.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
364 Litter mass decreased from spring to autumn in the treatments except for retention tree groups
365 (Fig. 4a). Litter pH increased in gap-cutting and clear-cutting in autumn, while in preparation
366 cutting and retention tree group it stayed close to the values measured in the control (Fig. 4b).
367 Litter moisture content increased marginally in preparation cutting and rather particularly in clear-
368 cutting as well as in gap-cutting as compared to the degree of humidity measured in spring
369 (Fig. 4c). We found that time effect was significant for pH and KAL-concentration (Table 2). Soil
370 pH was lower in spring than in autumn, in both periods, it was higher in retention tree group than
371 in other treatments (Fig. 4).
372
373 Diurnal pattern of microclimate variables among the treatments
374 When we analyzed the 24 hour datasets, a clear diurnal pattern could be detected for the different
375 microclimate variables with large variability between the treatments in summer. Contrarily, when
376 the pooled early spring and autumn subsets were analyzed, differences were much smaller and the
377 pattern was not that obvious (Fig. 5). In the case of light, we found that in summer, a large
378 difference came off among the treatments, the amount of PAR in clear-cutting could exceed 2000
379 μEm−2s−1, while in the second brightest plots, in gap-cutting, the maximum values were under
380 1930 μEm−2s−1. There was a detectable lag of the maximum values also: in clear-cutting at 12:00-
381 12:20 UTC, in gap-cutting at around 12:30-12:40 UTC and in preparation cutting at 13:10-13:20
382 UTC. In retention tree groups the values showed interesting pattern – presumably according to
383 their spatial configurations, i.e. the surrounding clear-cutting –, there is a certain excessive amount
384 of light from east, therefore the maximums occurred between 9:00 and 10:30 UTC (in the morning
385 there is an observable period when the light is higher in retention tree groups than in the other
386 treatments), and that was followed by a decrease in radiation because of the shading effect of the
387 canopy patch. The differing pattern of irradiance among the treatments was also detectable in the
388 case of VPD and soil temperature: e.g. in the morning, retention tree group was the warmest and
389 driest treatment. In clear-cutting, soil temperature could reach 38.8 °C in summer, but even in
390 retention tree group, Tsoil was maximized in 31.3 °C. In the gap-cutting, the moist soil (SWC was
391 the highest; Fig. 2) resulted a distinct peak in VPD, which was followed by a quick decrease, and
392 soil temperature was lower despite the higher total light than in preparation cutting. Clear-cutting
393 could cool down the most: in the summer, between 2:00 and 7:30 AM and especially in transitional
394 period, it was the coldest treatment. In the transition period (in the right panels of Fig. 5), the
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
395 amplitudes of the diurnal cycles are substantially smaller. Furthermore, the applied treatments do
396 not differ as much as during the peak of the growing season.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
397 DISCUSSION
398
399 Rapid changes in microclimate and litter variables, but not in soil properties
400 In general, microclimate variables and litter attributes showed strong short-term response among
401 the different silvicultural treatments, but we could not measure significant differences for most of
402 the investigated soil chemical variables.
403 Two different, but highly interrelated processes can be highlighted in the context of microclimate
404 alteration by forest management: radiation balance and evapotranspiration. Forest canopy plays an
405 important role in both mechanisms. The maintenance of the buffered microclimate in the below-
406 canopy space of closed forest stands is based on its shielding effect: through the (partial) shading
407 and the absorption of the foliage there is significantly less net radiation to heat the forest floor;
408 moreover, the canopy insulates the understory environment by reducing the longwave radiative
409 loss (Geiger, Aron & Todhunter, 1995; Rambo & North, 2009; von Arx et al., 2013). Furthermore,
410 as it was demonstrated by Bristow and Campbell (1984) there is a strong correlation between solar
411 irradiance and transferred heat-related variables of the ambient air such as air temperature, relative
412 humidity and vapor pressure deficit. Therefore, in general, in the harvested sites, we measured
413 higher and temporally more variable irradiance, temperatures, vapor pressure deficit and lower,
414 but also more unbalanced air humidity. Changes in soil moisture following the different treatments
415 are based (1) on the lower rate of interception and evaporation through the canopies or trunks and
416 consequently the increased throughfall; and (2) on lower transpiration rates due to tree removal
417 (Keenan & Kimmins, 1993; Wood, Hannah & Sadler, 2007; Chang, 2013; Muscolo et al., 2014).
418 The major effect of these alterations is that soil moisture typically increases in sites where felling
419 was applied on a larger continuous extent (i.e. gap-cutting and clear-cutting).
420
421 Light variables
422 As the applied management practices were planned based both on the total basal area and spatial
423 arrangement of the retained standing trees, we found that (1) total and diffuse light departed from
424 control in every treatment due to harvest-induced canopy modifications, and (2) the light variables
425 had the strongest response to the treatments. The amount and the range of light was the largest in
426 clear-cutting, and decreased in the following order: gap-cutting, retention tree group, preparation
427 cutting and control. Our findings are congruent with previous researches showing that the
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
428 increment in irradiance and its variability are larger as the size of canopy openings increase (or
429 even more as sky view factor enhances, Oke, 1987) that is highly correlated with the leaf area
430 index (Geiger, Aron & Todhunter, 1995; Carlson & Groot, 1997; Aussenac, 2000; Hardwick et
431 al., 2015). Therefore, solar radiation is higher and temporally more variable in clear-cutting than
432 in forest edges (Matlack, 1993), stands harvested by various types of green tree retention schemes
433 (Heithecker & Halpern, 2006) or the management practices related to the uneven-aged systems
434 (Zheng et al., 2000). This altered light regime with extreme means and maxima created by clear-
435 cutting is substantially different from any forest environment. Keenan and Kimmins (1993)
436 mentioned that it could be harmful to assimilating organs: in clear-felled areas the extreme
437 radiation (10- to 20-fold increase compared to closed forest stand) – and as a consequence, the
438 strikingly increased leaf surface temperatures – could suppress the efficiency of photosynthesis
439 and even destruct plant tissues. Canopy gaps also create a more illuminated environment (Gray,
440 Spies & Easter, 2002; Ritter, Dalsgaard & Einhorn, 2005; Abd Latif & Blackburn, 2010), though
441 the irradiance was significantly lower than it was detected in clear-cutting regarding the smaller
442 sky view factor (Carlson & Groot, 1997) and consequently, the shading of the surrounding trees
443 (Gálhidy et al., 2006). According to the light level and its variance, preparation cutting and
444 retention tree group did not show clear distinction, these treatments provide similarly brighter
445 environment than the uncut sites (Heithecker & Halpern, 2006; Brose, 2011; Grayson et al., 2012).
446 The spatial arrangement of the trees had influence on the direct-diffuse light proportions: in the
447 preparation cutting, the uniform distribution of trees could strongly inhibit the direct irradiation,
448 but less notably the diffuse light; therefore, the amount of the diffuse insolation is similar to that
449 in the gap-cutting, but the amount of total light is significantly lower (Abd Latif & Blackburn,
450 2010; Musselman, Pomeroy & Link, 2015). Retention tree group – in the first year – was very
451 open to the adjacent clear-cutting and to the lateral irradiance due to the lack of lower branches
452 and scarce shrub layer. However, we can expect that illumination in retention tree groups will
453 decrease and return to the level characteristic in control as the natural regeneration (shrubs, sprouts
454 and juvenile trees) grows and as the epicormics shoots emerge.
455
456 Air variables
457 Forest management (especially clear-cutting) has a long-lasting effect on air temperature and
458 relative humidity (Dovčiak & Brown, 2014; Baker et al., 2014); silvicultural treatments could
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
459 generate alterations in these variables that persist over 25 years. Contrary to previous studies
460 reporting that temperature and humidity in stands harvested by moderately intensive management
461 practices had only slight modifications on microclimate compared to uncut plots (e.g. group
462 selection or patch-cuts – Brooks & Kyker-Snowman, 2008; weak thinning – Weng et al., 2007;
463 gaps – Muscolo et al., 2014), we found that almost every treatment types resulted significant
464 departures from control in these variables. Our treatments resulted similar trends in Tair, RH and
465 VPD changes as other studies (e.g. Chen, Franklin & Spies, 1995; Gray, Spies & Easter, 2002;
466 Heithecker & Halpern, 2006; Ma et al., 2010).
467 Since the highest levels of incoming solar radiation were measurable in clear-cutting, this
468 treatment type can be characterized by the highest air temperature and vapor pressure deficit along
469 with the lowest relative humidity values (Keenan & Kimmins, 1993; Chen et al., 1999; Heithecker
470 & Halpern, 2006). Our findings are in agreement with the results of Carlson and Groot (1997) and
471 von Arx and colleagues (2012) as means of Tair in clear-cuttings were less than 1°C warmer than
472 in control plots. Differences in Tair or RH between clearings and closed stands were greater in
473 many studies that could be addressed to the larger clearing size. In our experiment, due to the size
474 of 0.5 ha of the clear-cutting, the shading effect of the forest edge as well as the cooling effect by
475 mixing air from the nearby stand could be more pronounced (Davies-Colley, Payne & van Elswijk,
476 2000; Baker, Jordan & Baker, 2016; Arroyo-Rodríguez et al., 2017).
477 Since the highest levels of incoming solar radiation were measurable in clear-cutting, this
478 treatment type can be characterized by the highest air temperature and vapor pressure deficit as
479 well as the lowest relative humidity values (Keenan & Kimmins, 1993; Chen et al., 1999;
480 Heithecker & Halpern, 2006). Our findings are in agreement with the results of Carlson and Groot
481 (1997) and von Arx et al. (2012) as mean of Tair in clear-cutting were less than 1°C warmer than
482 that of control plots. Differences in Tair or RH between clearings and closed stands were greater in
483 many studies that could be addressed to the larger clearing size. In our experiment, due to the
484 relatively small size (0.5 ha) of the clear-cutting, the shading effect of the forest edge as well as
485 the cooling effect by mixing air from the nearby stand could be more pronounced (Davies-Colley,
486 Payne & van Elswijk, 2000; Baker, Jordan & Baker, 2016; Arroyo-Rodríguez et al., 2017).
487 Regarding to the gap-cutting, despite the high PAR values and the generally observed strong
488 correlation between direct radiation and air temperature in gaps (Gray, Spies & Easter, 2002), in
489 our case the mean Tair and VPD in gaps were significantly lower than those in clear-cutting or
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
490 retention tree group. As an explanation the higher soil moisture in gap-cutting should be
491 underlined: as a consequence, energy absorption by the water particles in the air and evaporative
492 cooling is more pronounced in gaps (Robson et al., 2008; von Arx et al., 2013). In the applied gap-
493 cutting, the mean increment in temperature (~0.1°C) was comparable with the studies of Carlson
494 and Groot (1997) or Abd Latif and Blackburn (2010). Air humidity could remain unaltered in gaps
495 due to the lowered ratio of air mixing and the shading by adjacent closed stands (Geiger, Aron &
496 Todhunter, 1995; Abd Latif & Blackburn, 2010) and the moist topsoil as a source of water vapour
497 (Ogée & Brunet, 2002).
498 The thermal conditions of the preparation cutting (where 70% of the original BA was retained)
499 were similar to gap-cutting. The mean of VPD and RH in preparation cutting remained similar to
500 the control, but the ranges of these variables departed due to the higher input of solar energy (Weng
501 et al., 2007).
502 We expected that a retained patch of overstory trees in the clear-felled area could substantially
503 buffer the thermal effects of clear-cutting – as it was measured in case of light variables. However,
504 it was found that retention tree group could not compensate the thermal loading and the drying
505 capacity of the warmer air coming from the clearing: the mean of Tair and VPD were not
506 significantly different from those of the clear-cutting. This phenomenon could be addressed to the
507 edge effect (in the case of Tair, VPD and RH) that overlaps with the applied patch size (Matlack,
508 1993; Ewers and Banks-Leite, 2013; Baker et al, 2016; Arroyo-Rodriguez et al. 2017). In contrast,
509 retention tree group could successfully reduce the variability of the extreme values by the
510 insulation of the remained patch of canopies despite the effective lateral mixing (Heithecker &
511 Halpern, 2007; Ewers & Banks-Leite, 2013).
512
513 Soil temperature
514 Increased incoming solar radiation caused significant increment in soil temperature (both in the
515 case of means and IQRs), in all treatments. Higher departures from control levels were measured
516 than in the case of air temperature. As it was emphasized by von Arx et al. (2013), despite the
517 shading by the canopy affects both Tsoil and Tair temperature, the moderating effect is typically less
518 pronounced regarding air temperature. It is mainly because air is a mobile medium, thus turbulent
519 mixing of the air reduces the differences more (Morecroft, Taylor & Oliver, 1998; Hari, Heliövaara
520 & Kulmala, 2013). According to the strong correlation between Tair and Tsoil (Kang et al., 2000;
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
521 Ritter, Dalsgaard & Einhorn, 2005), the decreasing order of these variables across the treatments
522 was similar. Implicitly, because of the enhanced solar heating in daytime and the longwave
523 radiation loss in nighttime, clear-cutting was led to the greatest increase in Tsoil regarding both
524 mean and variability (Carlson & Groot, 1997; Davies-Colley, Payne & van Elswijk, 2000; von
525 Arx, Dobbertin & Rebetez, 2012). Retention tree group could moderate the extremes of Tsoil better
526 than Tair due to shading of the patch of standing trees. It is similar to the results of Williams-Linera,
527 Dominguez-Gastelu & Garcia-Zurita (1998) about isolated trees, but in our case a more
528 pronounced smoothing effects was recorded on the variability of Tsoil. Moreover, the range of Tsoil
529 in retention tree group was comparable with that measured in preparation cutting and gaps. For
530 gaps and preparation cutting, we recorded smaller increase in Tsoil than it was reported by previous
531 studies (Gray, Spies & Easter, 2002; Ritter, Dalsgaard & Einhorn, 2005; Abd Latif & Blackburn,
532 Thibodeau et al., 2000).
533
534 Soil moisture
535 Forest stands have high evapotranspiration rates; therefore, as a general rule, any opening in the
536 canopy cover results in a reduction in the amount of water consumed (Aussenac, 2000). The impact
537 of the elimination of tree individuals is particularly significant on soil moisture content (Keenan
538 & Kimmins, 1993; Gray, Spies & Easter, 2002). In clear-cutting, the soil moisture was
539 significantly higher than in the control, because of the drastic decrease of transpiring surface but
540 it was lower than in gap-cuttings due to the enhanced irradiation that increased the evaporation of
541 the surface and the great wind exposure (Geiger, Aron & Todhunter, 1995). The greatest increment
542 was detectable in gap-cutting as it was expected according to previous studies (Gray, Spies &
543 Easter, 2002; Ritter, Dalsgaard & Einhorn, 2005; Gálhidy et al., 2006; Abd Latif & Blackburn,
544 2010). This change of the water balance is usually rapid: initially soil moisture increases after the
545 applied treatments, but drops to the pre-harvest levels within a short time period following the
546 reestablishment of vegetation. This time-span is approximately four-five years in thinned stands
547 and clear-cuttings (Aussenac & Granier, 1988; Adams, Flint & Fredriksen, 1991). A similar, but
548 even faster return to the pre-harvest level of SWC could be predicted for? gaps as well
549 (Lewandowski et al., 2015), due to the development of the natural regeneration, the increased root
550 extraction and improved interception by the enhanced lateral growth of the surrounding trees
551 (Gray, Spies & Easter, 2002; Ritter, Dalsgaard & Einhorn, 2005). Here, we found that in retention
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
552 tree group, despite the higher VPD, the enhanced heat load and the transpiration of remnant trees,
553 soil water content was only slightly lower than in the uncut plots. SWC in the preparation cutting
554 remained also similar to the intact stands, although every third tree individual was cut. We can
555 suppose that this moderately increased amount of throughfall and somewhat lowered
556 evapotranspiration rate could not increase the water table as notably as it was suggested by
557 previous thinning studies (Aussenac & Granier, 1988; Chase et al., 2016), but this result is
558 comparable with the findings of Weng et al. (2007).
559
560 Litter variables
561 As the microclimate variables, litter characteristics showed a rapid response to the management
562 types. Albeit the amount of litter did not depart significantly from control and showed non-
563 significant treatment effect, there are observable differences between the treatment levels:
564 retention tree group could be characterized by litter accumulation due to the continuity of the local
565 tree litter input and lower soil moisture content that could slow the decomposition rates. This
566 phenomenon can also be explained by the different quality of the litter and the understory
567 vegetation. In clear-cutting and gap-cutting the cover of the herbaceous species considerably
568 increased (Tinya pers. comm.), that resulted in higher proportion of herb leaves in the litter mass.
569 Its decomposition provides more neutral litter conditions than the leaves of trees, especially leaves
570 of oak species (Finzi, Canham & Van Breemen, 1998). The higher litter moisture in clear-cutting
571 and gap-cutting can be explained by the higher SWC and the high understory cover that can
572 enhance humid conditions by insulating the surface (Keenan & Kimmins, 1993).
573
574 Soil chemical variables
575 Changes in nutrient availability and -cycling following forestry treatments are complex and in
576 numerous cases, trajectories are governed by multiple factors – thus studies show inconsistent results
577 (Nykvist & Rosén, 1985; Thiffault et al., 2011; Binkley & Fisher, 2013). In general, the regional
578 climate, the soil type and tree species composition can be emphasized as important determinants
579 (Keenan & Kimmins, 1993; Nave et al., 2010); and based on long-term data biomass removal per se
580 appears to have been little or no effect on site fertility, the effects are mostly transitory (Binkley &
581 Fisher, 2013). Nevertheless, nutrient loss following clear-cutting is typically reported (Lindo &
582 Visser, 2003) – especially in the case of available N (Jerabkova et al., 2011) –, while gaps are known
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
583 as environments with high rate of soil organic matter decomposition and mineralization causing
584 increased levels of nutrients (Muscolo et al., 2014). It is also suggested that nutrient losses can be
585 reduced by applying harvesting practices that only cause smaller scale disturbances such as gap-
586 cutting (Ritter, Starr & Vesterdal, 2005) or partial-cut harvesting (Lindo & Visser, 2003).
587 Based on the models we can conclude that only pH of the upper mineral soil showed immediate
588 treatment effect, while the other soil chemical variables remained similar to those of uncut sites.
589 These findings are roughly concurring with studies that besides the harvest effects investigate the
590 temporal changes in soil properties (Thiffault et al., 2011; Kishchuk et al., 2014). However, in their
591 meta-analysis Jerabkova et al. (2011) found that in deciduous forests as an impact on nitrate
592 concentration and N flux following clear-cutting a prompt and short-lived increment is typical.
593 The treatment effect in soil pH could be addressed to the changes in soil moisture: only retention
594 tree group differed both from control and the other treatments, where the lowest SWC and litter
595 moisture were measured. The relationship between the soil moisture and acidity has various
596 correlation strength (it is highly dependent e.g. from the soil type), but in general pH and SWC show
597 reverse pattern (Allen et al., 1989; Ji et al., 2014). In retention patches of deciduous forest stands,
598 Lando and Visser (2003) found also decreased level of soil acidity. As Sayer (2005) pointed out,
599 natural accumulation of litter could be associated with an increase in pH.
600
601 Distinct temporal patterns over the first growing season
602 By investigating the temporal pattern of the microclimate variables throughout the growing season
603 we found that (1) the differences between treatment levels increase as the shading capacity –
604 provided by the tree canopy – enhances; and (2) the evapotranspiration rates increase as the
605 vegetation becomes fully-leaved. This was unambiguous in the case of total light: during the
606 emergence and senescence phase the light conditions were homogeneous among the applied
607 treatments – except for the clear-cutting due to the lack of shading trunks and branches –, whereas
608 during the vegetation peak a very pronounced treatment effect could be detected. The seasonal
609 changes of soil temperature followed the pattern of the incoming radiation, except in gaps. In clear-
610 cutting, Tsoil was much lower in the end of the growing season than in the other treatments that
611 suggests the more pronounced frost-exposure of the regeneration (Aussenac, 2000; von Arx,
612 Dobbertin & Rebetez, 2012). In general, the buffering capacity of the forests concerning numerous
613 microclimate variables (e.g. the diurnal variability of Tair, RH, VPD or Tsoil) is related to the leaf area
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
614 index and also to the soil water potential (Ashcroft & Gollan, 2013; von Arx et al., 2013; Hardwick
615 et al., 2015). Gray et al. (2002) found similar annual pattern in the case of SWC means in gaps vs.
616 uncut control.
617 For the litter and soil variables, in the most cases, time had stronger effect than the treatment levels.
618 From spring to autumn – due to the reduced litter inputs and the enhanced decomposition rates –
619 litter mass decreased in most treatments. The only exception was the retention tree group, where the
620 dry and warm soil could deplete the performance of the decomposer organisms (Lindo & Visser,
621 2003). The role of microbial communities, fungi and soil fauna could be corroborated by the results
622 of Boros (unpublished data) who found that both the abundance and species richness of Enchytraeid
623 worms – an important decomposer group in the temperate zone (Schaefer, 1990) – decreased
624 significantly in retention tree groups. Litter pH and litter moisture content followed the similar
625 pattern as SWC, both variables showed a considerable increase in gap-cutting and clear-cutting from
626 spring to autumn. We can speculate that this increment can be the result of the enhanced leaching of
627 tannic acids, or the changed contribution of the different species to litter.
628 Soil pH increased in autumn in all treatments compared to that in control. That could be partly
629 explained by the weakened buffering capacity of litter layer on soil pH followed the decomposition
630 of leaf litter through the growing season (Sayer, 2005), although seasonal fluctuations in the range
631 of 0.5 unit is part of the natural dynamics.
632
633 Diurnal patterns across treatments differed more during the vegetation peak
634 According to the diurnal changes, clear differences were found between treatments that were more
635 pronounced during the fully-leaved period than in the transitional period. Moreover, the amplitudes
636 of the microclimate variables were also greater during the leaved period. The daily courses presented
637 here were mainly determined by the solar azimuth. During the transitional period (spring and
638 autumn) the maxima of the insolation were not influenced by canopy, thus it was synchronous in all
639 treatments. However, in summer, control was evenly shaded whole day, while light increment
640 peaked close to noon in the cclear-cutting. In the gap and preparation cutting some lags were
641 observable, because of the shading by the surrounding trees and the patchy environment,
642 respectively. In the retention tree group, a large proportion of total daily illuminance arrived in
643 forenoon – or before the sun passes through the meridian. This result confirms our hypothesis that
644 the main source of the total (and diffuse) light in the small forest remnants is the lateral irradiance.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
645 As it was suggested by previous studies (Carlson & Groot, 1997; Morecroft, Taylor & Oliver, 1998;
646 Holst, Mayer & Schindler, 2004), the turning points of the heat-driven variables – as VPD or soil
647 temperature presented here – are behind time to the pattern of PAR due to the latent heat loss.
648 However, von Arx, Dobbertin & Rebetez (2012) found that there is no significant time lag between
649 the daytime peaks of below-canopy and open-field temperature and relative humidity: according to
650 their interpretation, this quasi-synchronous pattern is a consequence to the main drivers of these two
651 variables: the solar radiation and vertical air exchange. By inspecting the VPD and Tsoil courses of
652 retention tree group we can detect the effect of enhanced insolation in the forenoon. According to
653 the temperature, we can speculate that the increase demonstrated via overall means is a consequence
654 of the elongated thermal load. It is also noticeable that VPD in gap-cutting increases synchronously
655 with that in the clear-cutting, but presumably because of enhanced evaporation of the moist soil
656 surface and the transpiration of herbs, its value sinks to the level of the uncut control (Ritter,
657 Dalsgaard & Einhorn, 2005; von Arx et al., 2013). As Aschroft and Gollan (2013) demonstrated
658 moister conditions reduce the diurnal variability of soil and air temperature and VPD more that could
659 be added as a further explanation for the more stable microclimate in gap-cutting.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
660 CONCLUSIONS AND MANAGEMENT IMPLICATIONS
661
662 Because of the relatively short time frame, the interpretation of our results has its limitations. The
663 ongoing measurements and the experimental setup give the opportunity to study the long-term
664 processes. We expect that (1) differences in microclimate variables between treatments are going
665 to weaken as the vegetation grows and variables will have different trajectories over time; (2) the
666 response of litter properties will be more pronounced and (3) soil variables will show treatment
667 effects in at least the cutting treatments. It is still a question how the different silvicultural
668 treatments create spatial pattern of abiotic variables. Therefore, our plan is to extend the
669 investigations to study the forest site conditions at fine-scale.
670 Using the results measured during the first growing season, we could demonstrate that the applied
671 management practices considerably changed the below-canopy microclimate in a short time
672 following the harvests. Clear-cutting had the most drastic impact on microclimate variables due to
673 the absence of tree canopies on large areas. According to the extreme light increment, the mean
674 air and soil temperature, vapor pressure deficit and their variability increased the most in this
675 treatment type. Organisms in clear-cutting are more exposed to thermal extremes and early frost
676 damages as well. Limited but positive moderating effect could be addressed to the application as
677 such small retention tree group even if the mean air and soil temperature and VPD are similar to
678 the clear-cutting. Gap-cutting provide more available light and consumable soil moisture that could
679 be favorable for herbs. Artificial gaps at this size can ensure the buffered environment. Preparation
680 cutting preserved forest conditions the most, although in Central Europe it is only a transitional
681 state before the terminal cutting.
682 We can conclude that in oak-hornbeam stands, for the achievement of the conservational aims and
683 to guarantee the more complete rate of ecosystem functionality, it is recommendable to apply
684 small-scale or spatially dispersed forestry treatments to preserve the original characteristics of the
685 forest environment as much as possible. Gap-cutting and – similarly to our preparation cutting –
686 irregular shelterwood or precommercial thinning can be suitable to achieve this aim. If the use of
687 the large, even-aged forestry practices is unavoidable, the application of different retention tree
688 group schemes seems to be particularly important (e.g. in clear-felled or slash and burn areas) to
689 provide the “lifeboat” environment for the forest-dwelling organism groups during the
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
690 regeneration (Heithecker & Halpern, 2007; Rosenvald & Lõhmus, 2008; Gustafsson, Kouki &
691 Sverdrup-Thygeson, 2010).
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
692 ACKNOWLEDGEMENTS
693 We are grateful for the cooperation and the joint efforts of the Pilisi Parkerdő Ltd., especially for
694 Péter Csépányi, Viktor Farkas, Gábor Szenthe and László Simon. We thank Tibor Standovár for
695 the LAI-2000 instruments. BK is deeply thankful for the help of Beáta Biri-Kovács, Blanka Biri
696 and András Guba.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
697 REFERENCES
698 Abd Latif Z., Blackburn GA. 2010. The effects of gap size on some microclimate variables during
699 late summer and autumn in a temperate broadleaved deciduous forest. International Journal of
700 Biometeorology 54:119–129. DOI: 10.1007/s00484-009-0260-1.
701 Act No. LVI of 2017 on forests, on the protection and management of forests. [2017. évi LVI.
702 törvény az erdőről, az erdő védelméről és az erdőgazdálkodásról szóló 2009. évi XXXVII.
703 törvény és egyéb kapcsolódó törvények módosításáról.] Magyar Közlöny 75: 7752–7796.
704 Adams PW., Flint AL., Fredriksen RL. 1991. Long-term patterns in soil moisture and revegetation
705 after a clearcut of a Douglas-fir forest in Oregon. Forest Ecology and Management 41:249–
706 263. DOI: 10.1016/0378-1127(91)90107-7.
707 Allen SD., Fish US., Research WS., Development, Science U of RID of NR. 1989. Soil-vegetation
708 correlations in transition zones of Rhode Island red maple swamps. U.S. Dept. of the Interior,
709 Fish and Wildlife Service, Research and Development.
710 Allen RG., Pereira LS., Raes D., Smith M. 1998. FAO Irrigation and drainage paper No. 56.
711 Rome.
712 Anderson DB. 1936. Relative Humidity or Vapor Pressure Deficit. Ecology 17:277–282. DOI:
713 10.2307/1931468.
714 Annighöfer P., Beckschäfer P., Vor T., Ammer C. 2015. Regeneration Patterns of European Oak
715 Species (Quercus petraea (Matt.) Liebl., Quercus robur L.) in Dependence of Environment and
716 Neighborhood. PLOS ONE 10:e0134935. DOI: 10.1371/journal.pone.0134935.
717 Arroyo-Rodríguez V., Saldaña-Vázquez RA., Fahrig L., Santos BA. 2017. Does forest
718 fragmentation cause an increase in forest temperature? Ecological Research 32:81–88. DOI:
719 10.1007/s11284-016-1411-6.
720 von Arx G., Dobbertin M., Rebetez M. 2012. Spatio-temporal effects of forest canopy on
721 understory microclimate in a long-term experiment in Switzerland. Agricultural and Forest
722 Meteorology 166–167:144–155. DOI: 10.1016/j.agrformet.2012.07.018.
723 von Arx G., Graf Pannatier E., Thimonier A., Rebetez M. 2013. Microclimate in forests with
724 varying leaf area index and soil moisture: potential implications for seedling establishment in a
725 changing climate. Journal of Ecology 101:1201–1213. DOI: 10.1111/1365-2745.12121.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
726 Ashcroft MB., Gollan JR. 2013. Moisture, thermal inertia, and the spatial distributions of near-
727 surface soil and air temperatures: Understanding factors that promote microrefugia.
728 Agricultural and Forest Meteorology 176:77–89. DOI: 10.1016/j.agrformet.2013.03.008.
729 Aussenac G. 2000. Interactions between forest stands and microclimate: Ecophysiological aspects
730 and consequences for silviculture. Annals of Forest Science 57:287–301. DOI:
731 10.1051/forest:2000119.
732 Aussenac G., Granier A. 1988. Effects of thinning on water stress and growth in Douglas-fir.
733 Canadian Journal of Forest Research 18:100–105. DOI: 10.1139/x88-015.
734 Baker TP., Jordan GJ., Baker SC. 2016. Microclimatic edge effects in a recently harvested forest:
735 Do remnant forest patches create the same impact as large forest areas? Forest Ecology and
736 Management 365:128–136. DOI: 10.1016/j.foreco.2016.01.022.
737 Baker TP., Jordan GJ., Steel EA., Fountain-Jones NM., Wardlaw TJ., Baker SC. 2014.
738 Microclimate through space and time: Microclimatic variation at the edge of regeneration
739 forests over daily, yearly and decadal time scales. Forest Ecology and Management 334:174–
740 184. DOI: 10.1016/j.foreco.2014.09.008.
741 Ballenegger R., di Gleria J. 1962. Methods of Soil and Fertilizer Analysis. Budapest,
742 Mezőgazdasdgi Kiadó.
743 Bartoń K. 2016. MuMIn: Multi-Model Inference.
744 Bernes C., Jonsson BG., Junninen K., Lõhmus A., Macdonald E., Müller J., Sandström J. 2015.
745 What is the impact of active management on biodiversity in boreal and temperate forests set
746 aside for conservation or restoration? A systematic map. Environmental Evidence 4. DOI:
747 10.1186/s13750-015-0050-7.
748 Bicknell JE., Struebig MJ., Edwards DP., Davies ZG. 2014. Improved timber harvest techniques
749 maintain biodiversity in tropical forests. Current Biology 24:R1119–R1120. DOI:
750 10.1016/j.cub.2014.10.067.
751 Bigelow SW., North MP. 2012. Microclimate effects of fuels-reduction and group-selection
752 silviculture: Implications for fire behavior in Sierran mixed-conifer forests. Forest Ecology and
753 Management 264:51–59. DOI: 10.1016/j.foreco.2011.09.031.
754 Binkley D., Fisher RF. 2013. Ecology and management of forest soils. Chichester: Wiley-
755 Blackwell.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
756 Bölöni J., Molnár Z., Biró M., Horváth F. 2008. Distribution of the (semi-)natural habitats in
757 Hungary II. Woodlands and shrublands. Acta Botanica Hungarica 50:107–148. DOI:
758 10.1556/ABot.50.2008.Suppl.6.
759 Bristow KL., Campbell GS. 1984. On the relationship between incoming solar radiation and daily
760 maximum and minimum temperature. Agricultural and Forest Meteorology 31:159–166. DOI:
761 10.1016/0168-1923(84)90017-0.
762 Brooks RT., Kyker-Snowman TD. 2008. Forest floor temperature and relative humidity following
763 timber harvesting in southern New England, USA. Forest Ecology and Management 254:65–
764 73. DOI: 10.1016/j.foreco.2007.07.028.
765 Brose PH. 2011. A comparison of the effects of different shelterwood harvest methods on the
766 survival and growth of acorn-origin oak seedlings. Canadian Journal of Forest Research
767 41:2359–2374. DOI: 10.1139/x11-143.
768 Brus DJ., Hengeveld GM., Walvoort DJJ., Goedhart PW., Heidema AH., Nabuurs GJ., Gunia K.
769 2012. Statistical mapping of tree species over Europe. European Journal of Forest Research
770 131:145–157. DOI: 10.1007/s10342-011-0513-5.
771 Carlson DW., Groot A. 1997. Microclimate of clear-cut, forest interior, and small openings in
772 trembling aspen forest. Agricultural and Forest Meteorology 87:313–329. DOI:
773 10.1016/S0168-1923(95)02305-4.
774 Carter MR., Gregorich EG. (eds.) 2008. Soil sampling and methods of analysis. [Pinawa,
775 Manitoba] : Boca Raton, FL: Canadian Society of Soil Science ; CRC Press.
776 Chang M. 2013. Forest hydrology: an introduction to water and forests. Boca Raton, FL: CRC
777 Press.
778 Chase CW., Kimsey MJ., Shaw TM., Coleman MD. 2016. The response of light, water, and
779 nutrient availability to pre-commercial thinning in dry inland Douglas-fir forests. Forest
780 Ecology and Management 363:98–109. DOI: 10.1016/j.foreco.2015.12.014.
781 Chaudhary A., Burivalova Z., Koh LP., Hellweg S. 2016. Impact of Forest Management on
782 Species Richness: Global Meta-Analysis and Economic Trade-Offs. Scientific Reports 6. DOI:
783 10.1038/srep23954.
784 Chen J., Franklin JF., Spies TA. 1995. Growing-Season Microclimatic Gradients from Clearcut
785 Edges into Old-Growth Douglas-Fir Forests. Ecological Applications 5:74–86. DOI:
786 10.2307/1942053.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
787 Chen J., Saunders SC., Crow TR., Naiman RJ., Brosofske KD., Mroz GD., Brookshire BL.,
788 Franklin JF. 1999. Microclimate in Forest Ecosystem and Landscape Ecology. BioScience
789 49:288–297. DOI: 10.2307/1313612.
790 Coulombe D., Sirois L., Paré D. 2017. Effect of harvest gap formation and thinning on soil nitrogen
791 cycling at the boreal–temperate interface. Canadian Journal of Forest Research 47:308–318.
792 DOI: 10.1139/cjfr-2016-0301.
793 Council Directive 92/43/EEC. Council Directive 92/43/EEC of 21 May 1992 on the conservation
794 of natural habitats and of wild fauna and flora.
795 Davies-Colley RJ., Payne GW., van Elswijk M. 2000. Microclimate gradients across a forest edge.
796 New Zealand Journal of Ecology 24:111–121.
797 De Frenne P., Rodriguez-Sanchez F., Coomes DA., Baeten L., Verstraeten G., Vellend M.,
798 Bernhardt-Romermann M., Brown CD., Brunet J., Cornelis J., Decocq GM., Dierschke H.,
799 Eriksson O., Gilliam FS., Hedl R., Heinken T., Hermy M., Hommel P., Jenkins MA., Kelly
800 DL., Kirby KJ., Mitchell FJG., Naaf T., Newman M., Peterken G., Petrik P., Schultz J., Sonnier
801 G., Van Calster H., Waller DM., Walther G-R., White PS., Woods KD., Wulf M., Graae BJ.,
802 Verheyen K. 2013. Microclimate moderates plant responses to macroclimate warming.
803 Proceedings of the National Academy of Sciences 110:18561–18565. DOI:
804 10.1073/pnas.1311190110.
805 DeForest JL., Chen J., McNulty SG. 2009. Leaf litter is an important mediator of soil respiration
806 in an oak-dominated forest. International Journal of Biometeorology 53:127–134. DOI:
807 10.1007/s00484-008-0195-y.
808 Dey DC. 2002. The ecological basis for oak silviculture in eastern North America. In: McShea
809 WJ, Healy WM eds. Oak forest ecosystems: ecology and management for wildlife. Baltimore:
810 Johns Hopkins University Press, 60–79.
811 Dieler J., Uhl E., Biber P., Müller J., Rötzer T., Pretzsch H. 2017. Effect of forest stand
812 management on species composition, structural diversity, and productivity in the temperate
813 zone of Europe. European Journal of Forest Research 136:739–766. DOI: 10.1007/s10342-
814 017-1056-1.
815 Dovčiak M., Brown J. 2014. Secondary edge effects in regenerating forest landscapes: vegetation
816 and microclimate patterns and their implications for management and conservation. New
817 Forests 45:733–744. DOI: 10.1007/s11056-014-9419-7.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
818 Dövényi Z. (ed.) 2010. Magyarország kistájainak katasztere [Hungarian microregions].
819 Budapest: MTA Földrajztudományi Kutatóintézet.
820 Ehbrecht M., Schall P., Ammer C., Seidel D. 2017. Quantifying stand structural complexity and
821 its relationship with forest management, tree species diversity and microclimate. Agricultural
822 and Forest Meteorology 242:1–9. DOI: 10.1016/j.agrformet.2017.04.012.
823 Ewers RM., Banks-Leite C. 2013. Fragmentation Impairs the Microclimate Buffering Effect of
824 Tropical Forests. PLoS ONE 8:e58093. DOI: 10.1371/journal.pone.0058093.
825 Faraway JJ. 2006. Extending the linear model with R: generalized linear, mixed effects and
826 nonparametric regression models. Boca Raton, Fla.: Chapman & Hall/CRC.
827 Finzi AC., Canham CD., Van Breemen N. 1998. Canopy tree–soil interactions within temperate
828 forests: species effects on ph and cations. Ecological Applications 8:447–454. DOI:
829 10.1890/1051-0761(1998)008[0447:CTSIWT]2.0.CO;2.
830 Frey SJK., Hadley AS., Johnson SL., Schulze M., Jones JA., Betts MG. 2016. Spatial models
831 reveal the microclimatic buffering capacity of old-growth forests. Science Advances
832 2:e1501392–e1501392. DOI: 10.1126/sciadv.1501392.
833 Furieri A. 2015. SpatiaLite.
834 Gálhidy L., Mihók B., Hagyó A., Rajkai K., Standovár T. 2006. Effects of gap size and associated
835 changes in light and soil moisture on the understorey vegetation of a Hungarian beech forest.
836 Plant Ecology 183:133–145. DOI: 10.1007/s11258-005-9012-4.
837 Gebauer T., Horna V., Leuschner C. 2012. Canopy transpiration of pure and mixed forest stands
838 with variable abundance of European beech. Journal of Hydrology 442–443:2–14. DOI:
839 10.1016/j.jhydrol.2012.03.009.
840 Geiger R., Aron RH., Todhunter P. 1995. The Climate Near the Ground. Wiesbaden:
841 Vieweg+Teubner Verlag. DOI: 10.1007/978-3-322-86582-3.
842 Gray AN., Spies TA., Easter MJ. 2002. Microclimatic and soil moisture responses to gap formation
843 in coastal Douglas-fir forests. Canadian Journal of Forest Research 32:332–343. DOI:
844 10.1139/x01-200.
845 Grayson SF., Buckley DS., Henning JG., Schweitzer CJ., Gottschalk KW., Loftis DL. 2012.
846 Understory light regimes following silvicultural treatments in central hardwood forests in
847 Kentucky, USA. Forest Ecology and Management 279:66–76. DOI:
848 10.1016/j.foreco.2012.05.017.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
849 Greiser C., Meineri E., Luoto M., Ehrlén J., Hylander K. 2018. Monthly microclimate models in
850 a managed boreal forest landscape. Agricultural and Forest Meteorology 250–251:147–158.
851 DOI: 10.1016/j.agrformet.2017.12.252.
852 Gustafsson L., Kouki J., Sverdrup-Thygeson A. 2010. Tree retention as a conservation measure in
853 clear-cut forests of northern Europe: a review of ecological consequences. Scandinavian
854 Journal of Forest Research 25:295–308. DOI: 10.1080/02827581.2010.497495.
855 Hardwick SR., Toumi R., Pfeifer M., Turner EC., Nilus R., Ewers RM. 2015. The relationship
856 between leaf area index and microclimate in tropical forest and oil palm plantation: Forest
857 disturbance drives changes in microclimate. Agricultural and Forest Meteorology 201:187–
858 195. DOI: 10.1016/j.agrformet.2014.11.010.
859 Hari P., Heliövaara K., Kulmala L. (eds.) 2013. Physical and physiological forest ecology.
860 Heidelberg ; London: Springer.
861 Heithecker TD., Halpern CB. 2006. Variation in microclimate associated with dispersed-retention
862 harvests in coniferous forests of western Washington. Forest Ecology and Management
863 226:60–71. DOI: 10.1016/j.foreco.2006.01.024.
864 Heithecker TD., Halpern CB. 2007. Edge-related gradients in microclimate in forest aggregates
865 following structural retention harvests in western Washington. Forest Ecology and
866 Management 248:163–173. DOI: 10.1016/j.foreco.2007.05.003.
867 Holst T., Mayer H., Schindler D. 2004. Microclimate within beech stands - part II: thermal
868 conditions. European Journal of Forest Research 123:13–28. DOI: 10.1007/s10342-004-0019-
869 5.
870 Hothorn T., Bretz F., Westfall P. 2008. Simultaneous Inference in General Parametric Models.
871 Biometrical Journal 50:346–363. DOI: 10.1002/bimj.200810425.
872 Ibarra JT., Martin M., Cockle KL., Martin K. 2017. Maintaining ecosystem resilience: functional
873 responses of tree cavity nesters to logging in temperate forests of the Americas. Scientific
874 Reports 7. DOI: 10.1038/s41598-017-04733-2.
875 ISO 10694:1995. Soil Quality – Determination of organic and total carbon after dry combustion
876 (elementary analysis). Geneva, Switzerland: International Organization for Standardization.
877 ISO 13878:1998. Soil Quality – Determination of total nitrogen content by dry combustion
878 (“elemental analysis”). Geneva, Switzerland: International Organization for Standardization.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
879 Jerabkova L., Prescott CE., Titus BD., Hope GD., Walters MB. 2011. A meta-analysis of the
880 effects of clearcut and variable-retention harvesting on soil nitrogen fluxes in boreal and
881 temperate forests. Canadian Journal of Forest Research 41:1852–1870. DOI: 10.1139/x11-
882 087.
883 Ji C-J., Yang Y-H., Han W-X., He Y-F., Smith J., Smith P. 2014. Climatic and Edaphic Controls
884 on Soil pH in Alpine Grasslands on the Tibetan Plateau, China: A Quantitative Analysis.
885 Pedosphere 24:39–44. DOI: 10.1016/S1002-0160(13)60078-8.
886 Kang S., Kim S., Oh S., Lee D. 2000. Predicting spatial and temporal patterns of soil temperature
887 based on topography, surface cover and air temperature. Forest Ecology and Management
888 136:173–184. DOI: 10.1016/S0378-1127(99)00290-X.
889 Keenan RJ., Kimmins JP (Hamish). 1993. The ecological effects of clear-cutting. Environmental
890 Reviews 1:121–144. DOI: 10.1139/a93-010.
891 Kishchuk BE., Quideau S., Wang Y., Prescott C. 2014. Long-term soil response to variable-
892 retention harvesting in the EMEND (Ecosystem Management Emulating Natural Disturbance)
893 experiment, northwestern Alberta. Canadian Journal of Soil Science 94:263–279. DOI:
894 10.4141/cjss2013-034.
895 Knapp BO., Olson MG., Larsen DR., Kabrick JM., Jensen RG. 2014. Missouri Ozark Forest
896 Ecosystem Project: A Long-Term, Landscape-Scale, Collaborative Forest Management
897 Research Project. Journal of Forestry 112:513–524. DOI: 10.5849/jof.13-102.
898 Kovács B., Tinya F., Ódor P. 2017. Stand structural drivers of microclimate in mature temperate
899 mixed forests. Agricultural and Forest Meteorology 234–235:11–21. DOI:
900 10.1016/j.agrformet.2016.11.268.
901 Krasilnikov P., Marti JJII., Arnold R., Shoba S. 2009. A Handbook of Soil Terminology,
902 Correlation and Classification. London, Sterling VA: Earthscan.
903 Latimer CE., Zuckerberg B. 2017. Forest fragmentation alters winter microclimates and
904 microrefugia in human-modified landscapes. Ecography 40:158–170. DOI:
905 10.1111/ecog.02551.
906 Lemmon PE. 1956. A Spherical Densiometer For Estimating Forest Overstory Density. Forest
907 Science 2:314–320. DOI: 10.1093/forestscience/2.4.314.
908 Lenoir J., Hattab T., Pierre G. 2017. Climatic microrefugia under anthropogenic climate change:
909 implications for species redistribution. Ecography 40:253–266. DOI: 10.1111/ecog.02788.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
910 Lenth RV. 2016. Least-Squares Means: The R Package lsmeans. Journal of Statistical Software
911 69. DOI: 10.18637/jss.v069.i01.
912 Lewandowski TE., Forrester JA., Mladenoff DJ., Stoffel JL., Gower ST., D’Amato AW., Balser
913 TC. 2015. Soil microbial community response and recovery following group selection harvest:
914 Temporal patterns from an experimental harvest in a US northern hardwood forest. Forest
915 Ecology and Management 340:82–94. DOI: 10.1016/j.foreco.2014.12.012.
916 LI-COR Inc. 1992. LAI-2000 Plant Canopy Analyzer. Instruction manual.
917 Lin H., Chen Y., Song Q., Fu P., Cleverly J., Magliulo V., Law BE., Gough CM., Hörtnagl L., Di
918 Gennaro F., Matteucci G., Montagnani L., Duce P., Shao C., Kato T., Bonal D., Paul-Limoges
919 E., Beringer J., Grace J., Fan Z. 2017. Quantifying deforestation and forest degradation with
920 thermal response. Science of The Total Environment 607–608:1286–1292. DOI:
921 10.1016/j.scitotenv.2017.07.062.
922 Lindo Z., Visser S. 2003. Microbial biomass, nitrogen and phosphorus mineralization, and
923 mesofauna in boreal conifer and deciduous forest floors following partial and clear-cut
924 harvesting. Canadian Journal of Forest Research 33:1610–1620. DOI: 10.1139/x03-080.
925 Lõhmus A., Lõhmus P. 2010. Epiphyte communities on the trunks of retention trees stabilise in
926 5years after timber harvesting, but remain threatened due to tree loss. Biological Conservation
927 143:891–898. DOI: 10.1016/j.biocon.2009.12.036.
928 Ma S., Concilio A., Oakley B., North M., Chen J. 2010. Spatial variability in microclimate in a
929 mixed-conifer forest before and after thinning and burning treatments. Forest Ecology and
930 Management 259:904–915. DOI: 10.1016/j.foreco.2009.11.030.
931 Marshall V. 2000. Impacts of forest harvesting on biological processes in northern forest soils.
932 Forest Ecology and Management 133:43–60. DOI: 10.1016/S0378-1127(99)00297-2.
933 Matlack GR. 1993. Microenvironment variation within and among forest edge sites in the eastern
934 United States. Biological Conservation 66:185–194. DOI: 10.1016/0006-3207(93)90004-K.
935 Matthews JD. 1991. Silvicultural Systems. Clarendon Press.
936 Morecroft M., Taylor M., Oliver H. 1998. Air and soil microclimates of deciduous woodland
937 compared to an open site. Agricultural and Forest Meteorology 90:141–156. DOI:
938 10.1016/S0168-1923(97)00070-1.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
939 MSZ 20135:1999. A talaj oldható tápelemtartalmának meghatározása. [Determination of the
940 soluble nutrient element content of the soil]. Budapest: Hungarian Standards Institution
941 (MSZT).
942 MSZ-08-0205:1978. A talaj fizikai és vízgazdálkodási tulajdonságainak vizsgálata.
943 [Determination of physical and hydrophysical properties of soils]. Budapest: Hungarian
944 Standards Institution (MSZT).
945 Muscolo A., Bagnato S., Sidari M., Mercurio R. 2014. A review of the roles of forest canopy gaps.
946 Journal of Forestry Research 25:725–736. DOI: 10.1007/s11676-014-0521-7.
947 Musselman KN., Pomeroy JW., Link TE. 2015. Variability in shortwave irradiance caused by
948 forest gaps: Measurements, modelling, and implications for snow energetics. Agricultural and
949 Forest Meteorology 207:69–82. DOI: 10.1016/j.agrformet.2015.03.014.
950 Nave LE., Vance ED., Swanston CW., Curtis PS. 2010. Harvest impacts on soil carbon storage in
951 temperate forests. Forest Ecology and Management 259:857–866. DOI:
952 10.1016/j.foreco.2009.12.009.
953 Nykvist N., Rosén K. 1985. Effect of clear-felling and slash removal on the acidity of Northern
954 coniferous soils. Forest Ecology and Management 11:157–169. DOI: 10.1016/0378-
955 1127(85)90024-6.
956 Ogée J., Brunet Y. 2002. A forest floor model for heat and moisture including a litter layer. Journal
957 of Hydrology 255:212–233. DOI: 10.1016/S0022-1694(01)00515-7.
958 Oke TR. 1987. Boundary Layer Climates. London: Methuen.
959 Paillet Y., BergèS L., HjäLtéN J., óDor P., Avon C., Bernhardt-RöMermann M., Bijlsma R-J., De
960 Bruyn L., Fuhr M., Grandin U., Kanka R., Lundin L., Luque S., Magura T., Matesanz S.,
961 MéSzáRos I., Sebastià M-T., Schmidt W., StandováR T., TóThméRéSz B., Uotila A.,
962 Valladares F., Vellak K., Virtanen R. 2010. Biodiversity Differences between Managed and
963 Unmanaged Forests: Meta-Analysis of Species Richness in Europe. Conservation Biology
964 24:101–112. DOI: 10.1111/j.1523-1739.2009.01399.x.
965 Pinheiro J., Bates D., DebRoy S., Sarkar D., R Core Team. 2017. nlme: Linear and Nonlinear
966 Mixed Effects Models.
967 R Core Team. 2017. R: A language and environment for statistical computing. Vienna, Austria: R
968 Foundation for Statistical Computing.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
969 Rambo TR., North MP. 2009. Canopy microclimate response to pattern and density of thinning in
970 a Sierra Nevada forest. Forest Ecology and Management 257:435–442. DOI:
971 10.1016/j.foreco.2008.09.029.
972 Reiczigel J., Harnos A., Solymosi N. 2014. Biostatisztika nem statisztikusoknak. Nagykovácsi:
973 Pars Kft.
974 Ritter E., Dalsgaard L., Einhorn KS. 2005. Light, temperature and soil moisture regimes following
975 gap formation in a semi-natural beech-dominated forest in Denmark. Forest Ecology and
976 Management 206:15–33. DOI: 10.1016/j.foreco.2004.08.011.
977 Ritter E., Starr M., Vesterdal L. 2005. Losses of nitrate from gaps of different sizes in a managed
978 beech ( Fagus sylvatica ) forest. Canadian Journal of Forest Research 35:308–319. DOI:
979 10.1139/x04-185.
980 Robson TM., Rodriguez-Calcerrada J., Sanchez-Gomez D., Aranda I. 2008. Summer drought
981 impedes beech seedling performance more in a sub-Mediterranean forest understory than in
982 small gaps. Tree Physiology 29:249–259. DOI: 10.1093/treephys/tpn023.
983 Rosenvald R., Lõhmus A. 2008. For what, when, and where is green-tree retention better than
984 clear-cutting? A review of the biodiversity aspects. Forest Ecology and Management 255:1–
985 15. DOI: 10.1016/j.foreco.2007.09.016.
986 Runkle JR. 1981. Gap Regeneration in Some Old-growth Forests of the Eastern United States.
987 Ecology 62:1041–1051. DOI: 10.2307/1937003.
988 Ryu S-R., Concilio A., Chen J., North M., Ma S. 2009. Prescribed burning and mechanical thinning
989 effects on belowground conditions and soil respiration in a mixed-conifer forest, California.
990 Forest Ecology and Management 257:1324–1332. DOI: 10.1016/j.foreco.2008.11.033.
991 Sayer EJ. 2005. Using experimental manipulation to assess the roles of leaf litter in the functioning
992 of forest ecosystems. Biological Reviews 81:1–31. DOI: 10.1017/S1464793105006846.
993 Schaefer M. 1990. The soil fauna of a beech forest on limestone: trophic structure and energy
994 budget. Oecologia 82:128–136. DOI: 10.1007/BF00318544.
995 Schall P., Gossner MM., Heinrichs S., Fischer M., Boch S., Prati D., Jung K., Baumgartner V.,
996 Blaser S., Böhm S., Buscot F., Daniel R., Goldmann K., Kaiser K., Kahl T., Lange M., Müller
997 J., Overmann J., Renner SC., Schulze E-D., Sikorski J., Tschapka M., Türke M., Weisser WW.,
998 Wemheuer B., Wubet T., Ammer C. 2018. The impact of even-aged and uneven-aged forest
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
999 management on regional biodiversity of multiple taxa in European beech forests. Journal of
1000 Applied Ecology 55:267–278. DOI: 10.1111/1365-2664.12950.
1001 Seibold S., Bässler C., Brandl R., Gossner MM., Thorn S., Ulyshen MD., Müller J. 2015.
1002 Experimental studies of dead-wood biodiversity — A review identifying global gaps in
1003 knowledge. Biological Conservation 191:139–149. DOI: 10.1016/j.biocon.2015.06.006.
1004 Standovár T. 2006. Biológiai megfontolások az erdei életközösségek hatékony védelméhez.
1005 Magyar Tudomány 113:656–662.
1006 Stewart-Oaten A., Murdoch WW., Parker KR. 1986. Environmental Impact Assessment:
1007 “Pseudoreplication” in Time? Ecology 67:929–940. DOI: 10.2307/1939815.
1008 Stoffel JL., Gower ST., Forrester JA., Mladenoff DJ. 2010. Effects of winter selective tree harvest
1009 on soil microclimate and surface CO2 flux of a northern hardwood forest. Forest Ecology and
1010 Management 259:257–265. DOI: 10.1016/j.foreco.2009.10.004.
1011 Suggitt AJ., Gillingham PK., Hill JK., Huntley B., Kunin WE., Roy DB., Thomas CD. 2011.
1012 Habitat microclimates drive fine-scale variation in extreme temperatures. Oikos 120:1–8. DOI:
1013 10.1111/j.1600-0706.2010.18270.x.
1014 Thibodeau L., Raymond P., Camiré C., Munson AD. 2000. Impact of precommercial thinning in
1015 balsam fir stands on soil nitrogen dynamics, microbial biomass, decomposition, and foliar
1016 nutrition. Canadian Journal of Forest Research 30:229–238. DOI: 10.1139/x99-202.
1017 Thiffault E., Hannam KD., Paré D., Titus BD., Hazlett PW., Maynard DG., Brais S. 2011. Effects
1018 of forest biomass harvesting on soil productivity in boreal and temperate forests — A review.
1019 Environmental Reviews 19:278–309. DOI: 10.1139/a11-009.
1020 Verstraeten LMJ., Livens J. 1971. Hygroscopicity as a valuable complement in soil analysis, 1.
1021 Characterization of the hygroscopic constant. Geoderma 6:255–262. DOI: 10.1016/0016-
1022 7061(71)90057-7.
1023 Weng S-H., Kuo S-R., Guan BT., Chang T-Y., Hsu H-W., Shen C-W. 2007. Microclimatic
1024 responses to different thinning intensities in a Japanese cedar plantation of northern Taiwan.
1025 Forest Ecology and Management 241:91–100. DOI: 10.1016/j.foreco.2006.12.027.
1026 Wijdeven S., van Hees A. 2001. Natural gaps in beech forests in The Netherlands, North-West
1027 Germany, Belgium and France. In: Natural canopy gap characteristics in European beech
1028 forests. Nat-Man Working Report 2. Forest & Landscape Denmark, 29.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
1029 Williams-Linera G., Dominguez-Gastelu V., Garcia-Zurita ME. 1998. Microenvironment and
1030 Floristics of Different Edges in a Fragmented Tropical Rainforest. Conservation Biology
1031 12:1091–1102. DOI: 10.1046/j.1523-1739.1998.97262.x.
1032 Wood PJ., Hannah DM., Sadler JP. (eds.) 2007. Hydroecology and ecohydrology: past, present
1033 and future. Chichester, England ; Hoboken, NJ: Wiley.
1034 Zheng D., Chen J., Song B., Xu M., Sneed P., Jensen R. 2000. Effects of silvicultural treatments
1035 on summer forest microclimate in southeastern Missouri Ozarks. Climate Research 15:45–59.
1036 Zuur AF. (ed.) 2009. Mixed effects models and extensions in ecology with R. New York, NY:
1037 Springer.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Figure 1
The study site of the “Pilis Experiment” in northern Hungary.
A) Site location (47°40’N, 18°54’E) in the Pilis Mountains. B) Experimental design showing the
five treatments replicated within six blocks. C) Aerial photograph revealing a block with the
applied forestry treatments: clear-cutting (red) with a retention tree group (blue);
preparation cutting (orange), gap-cutting (purple) and control (green). Drone photo by Viktor
Tóth.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Table 1(on next page)
Characteristics of forest structure around the plots before and after treatments.
Structural attributes (mean ± standard deviation) presented here are diameter at breast
height (DBH, cm), canopy height (m), basal area (m2ha-1) and canopy closure (%). Canopy
closure was measured by spherical densitometer Model-A (Lemmon, 1956). Letter ‘U’ refers
to upper layer and ‘S’ to sub-canopy layer. ‘C’ – control; ‘CC’ – clear-cutting; ‘G’ – gap-cutting;
‘P’ – preparation cutting and ‘R’ – retention tree group. Mean and SD were calculated based
on the six replicates for each treatment type. In the case of clear-cutting and preparation
cutting 0.5 ha sized sampling area (diameter=80 m) were used for forest structure
measurements, while for estimations in control, gap cutting and retention tree groups 0.03
ha area (diameter=20 m) were applied.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Pre-treatment (2014) Post-treatment (2015)DBH Height
Basal area Basal areaTreat
-ment
U S U S U S
Canopy
closure U S
Canopy
closure
C 28.0(±5.8) 11.9 (±3.8) 20.9(±1.5) 10.8(±3.5) 29.32(±0.12) 8.83(±0.10) 89.8(±2.6) 29.32(±0.12) 8.83(±0.10) 93.5(±3.9)
CC 28.0 (±5.7) 11.8(±4.2) 21.6(±1.6) 10.4(±3.8) 29.58(±6.47) 9.98(±4.66) 87.9(±3.6) 0.00 0.00 2.5(±2.1)
G 27.3(±5.3) 12.5(±2.8) 20.5(±1.1) 11.2(±2.9) 29.53(±9.03) 9.33(±4.51) 88.4(±4.4) 0.00 0.00 44.8(±10.4)
P 27.2 (±5.3) 10.9(±4.1) 21.2(±1.4) 10.0(±3.5) 28.07(±2.10) 8.03(±1.33) 89.4(±4.4) 19.67(±1.48) 0.00 70.2(±6. 9)
R 27.3 (±5.8) 11.1(±3.4) 20.4(±1.9) 11.8(±3.9) 30.47(±3.73) 8.17(±2.35) 88.7(±3.2) 30.47(±3.73) 8.17(±2.35) 81.9(±9.2)
1
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Table 2(on next page)
The results of linear mixed models performed for site condition variables.
PAR: photosynthetically active radiation (μEm−2s−1); DIFN: relative diffuse light (%);Tair: air
temperature (°C); RH: relative humidity (%); VPD: vapor pressure deficit (kPa); Tsoil: soil
temperature (°C); SWC: soil moisture (m3/m3); Litter mass: total mass of collected litter on
the surface (gm-2); Litter pH: litter pH in water; Litter moisture content: gravimetric moisture
content of litter samples (%); Soil pH: soil pH in water; hy: Kuron’s hygrscopicity (%); [SOC]:
total soil carbon content (%); [N]: total nitrogen content (%); [PAL]: concentration of AL-
soluble phosphorus (mg/100 g soil); [KAL]: concentration of AL-soluble potassium (mg/100 g
soil). ‘d’ refers to the difference from values measured in ‘Control’ plots. For modeling, 24-
hour-means were used except in the case of PAR, where daytime (6:00-18:00 UTC) means
were calculated. Treatment types: ‘CC’ – clear-cutting; ‘G’ – gap-cutting; ‘P’ – preparation
cutting and ‘R’ – retention tree group. Superscripts refer to significant differences among
treatments (pairwise Tukey comparisons, alpha=0.05), treatment codes marked with bold
indicates significant departures from control (alpha=0.05).
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Model Treatment Time Treatment:TimeDependent
variable Chi2
1p R2
LRF p F p F p
dPAR mean 454.711 <0.0001 0.922 225.579 <0.0001 133.928 <0.0001 8.941 <0.0001
dPAR IQR 343.698 <0.0001 0.852 114.259 <0.0001 57.575 <0.0001 6.292 <0.0001
dDIFN 29.086 <0.0001 0.766 21.699 <0.0001 -- -- -- --
dTair mean 273.305 <0.0001 0.781 21.888 <0.0001 54.082 <0.0001 4.903 <0.0001
dTair IQR 265.160 <0.0001 0.771 44.487 <0.0001 47.139 <0.0001 2.016 0.0086
dRH mean 46.096 <0.0001 0.434 5.177 0.0021 2.939 0.0105 0.609 0.8866
dRH IQR 125.451 <0.0001 0.569 14.054 <0.0001 16.694 <0.0001 1.275 0.2173
dVPD mean 122.668 <0.0001 0.595 13.2782 <0.0001 13.9286 <0.0001 1.8528 0.0267
dVPD IQR 259.555 <0.0001 0.823 37.279 <0.0001 63.435 <0.0001 5.491 <0.0001
dTsoil mean 261.975 <0.0001 0.768 9.107 <0.0001 44.611 <0.0001 7.368 <0.0001
dTsoil IQR 201.537 <0.0001 0.674 24.397 <0.0001 24.166 <0.0001 3.248 <0.0001
dSWC mean 109.965 <0.0001 0.534 29.145 <0.0001 2.3129 0.0292 1.089 0.3666
dLitter mass 21.338 0.0033 0.424 2.164 0.1097 10.812 0.0057 1.955 0.1387
dLitter pH 35.390 <0.0001 0.524 8.888 0.0002 8.685 0.0057 3.646 0.0218
dLitter
moisture
47.003 <0.0001 0.624 9.318 0.0001 16.478 0.0003 7.355 0.0009
dSoil pH 23.863 0.0012 0.544 3.633 0.0221 15.754 0.0003 0.041 0.9889
dhy 10.428 0.1656 0.219 2.824 0.0528 0.115 0.7369 0.426 0.7358
d[SOC] 5.008 0.6590 0.352 1.202 0.3242 0.159 0.6930 0.223 0.8799
d[N] 3.415 0.8442 0.357 0.912 0.4451 0.008 0.9316 0.074 0.9738
d[PAL] 10.308 0.1718 0.388 1.936 0.1418 1.034 0.3163 0.965 0.4200
d[KAL] 12.735 0.0788 0.299 1.641 0.1821 6.956 0.0124 0.173 0.9143
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
1
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Figure 2(on next page)
Changes in means and interquartile ranges (IQR) of relative values of site condition
variables among forestry treatments in Pilis Mountains, Hungary in 2015.
Treatment types are coded as ‘CC’ – clear-cutting; ‘G’ – gap-cutting; ‘P’ – preparation cutting and ‘R’ –
retention tree group. PAR: photosynthetically active radiation (μEm−2s−1); DIFN: relative diffuse light (%);
SWC: soil moisture (m3/m3); Tair: air temperature (°C); RH: relative humidity (%); VPD: vapor pressure deficit
(kPa); Tsoil: soil temperature (°C); Litter mass: total mass of collected litter on the surface (g/m2); Litter pH:
litter pH in water; Litter moisture content: gravimetric moisture content of litter samples (%); Soil pH: soil pH
in water. Letter ‘d’ in the variable abbreviations refers to the differences from the mean values measured in
the ‘Control’ plots. Full circles show the mean; vertical lines denote the standard deviation of the samples.
Letters designate the significant differences among treatments (pairwise comparisons based on the linear
mixed models; Tukey-test, alpha=0.05), while asterisks denote significant differences from values measured
at control plots (random intercept models, alpha=0.05). The horizontal blue line shows the level of control.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
0.00
0.10
0.200.25
0.15
0.05dVPD
IQR (
kPa)
a*
b*bc*
c*
Treatments
0200
400
600
800
1000
dPAR
IQR (μE
m−
2s−
1)
a*
b*
c*c*
20
40
60
80
0
dD
IFN
(%
)
a*
b*bc* c*
0.0
0.2
0.4
0.6
dTair
mean (
°C)
a*
b*b*
a*
dTair
IQR (
°C)
a*
b*bc*
c*
-2-1
01
2
dRH
mean (
%)
a*b b
ab
-0.05
0.00
0.05
0.10
dSW
Cm
ean (
m3/m
3)
a*
b*
acc
0.00
0.05
0.10
dVPD
mean (
kPa)
a*
bb
a*
02
46
8
dTsoil
IQR (
°C)
a*
b* b*
b*
CC G P R CC G P R CC G P R
-10
12
34
dTsoil
mean (
°C)
a*
b* b*
ab*
d L
itte
r m
ass (
g/m
2)
-400
-200
0200
400
aa
a a
d L
itte
r pH
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
a* a*
b b
d L
itte
r m
ois
ture
(%
)
-50
510
15
20
25
a*
a*
a*
b
d S
oil p
H
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
a a a
b*
CC G P R
a)
i) l)
g)e)
j)
p)o)n)m)
f)
b) c) d)
h)
k)
200
400
600
800
dPAR
mean (μE
m−
2s−
1)
a*
b*
c* c*
0
dRH
IQR (
%)
a*
b* bc*
ac*
01
23
4
0.0
0.5
1.0
1.5
2.0
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Figure 3(on next page)
Temporal variability of selected microclimate variables among experimental treatments
and months in Pilis Mountains, Hungary in 2015.
Treatment types are coded as ‘CC’ – clear-cutting; ‘G’ – gap-cutting; ‘P’ – preparation cutting
and ‘R’ – retention tree group. PAR: photosynthetically active radiation (μEm−2s−1); VPD:
vapor pressure deficit (kPa); Tsoil: soil temperature (°C) and SWC: soil moisture (m3/ m3).
Letter ‘d’ in the variable abbreviations refers to the differences from the mean values
measured in the ‘Control’ plots. Full circles show the mean; vertical lines denote the standard
deviation of the samples. Letters designate the significant differences among treatments and
months (pairwise comparisons based on the linear mixed models: Tukey-test, alpha=0.05).
The horizontal blue line shows the level of control.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
-0.0
50
.00
0.0
50
.10
0.1
50
.20
CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R
March April May June July August September October
Treatment
(mean±
SD
; kPa)
dVPD
NA
a a a a a a a a a a a a a b b ab a b bc ac a b bc ac a a a a a a a
a a bcab c bc a
CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R
March April May June July August September October
Treatment
(mean±
SD
; °C)
dT
soil
0-6
-4-2
24
6 a a a a a a a a a ab bc c a b b b a b bc ac a b ab a a a a a a b b b
a ab cc c bc ab d
CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R
March April May June July August September October
(mean±
SD
; m
3/m
3)
dSW
C
-0.1
0.0
0.1
0.2
Treatment
a a a a a a a a a a a a a b a a a b ac c ab a ab b ab a b b a a a a
abc ab cabc bc ab abc a
02
00
40
06
00
80
01
00
0
CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R CC G P R
March April May June July August September October
Treatment
(mean±
SD
; μE
m−
2s−
1)
dPAR
a b b b a b bc c a b c d a b c c a b c c a bc b c a b b b ac b ab c
a ab cc c c ab ba) b)
c) d)
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Figure 4(on next page)
Seasonal changes in selected soil and litter variables in Pilis Mountains, Hungary in
2015.
Treatment types are coded as ‘CC’ – clear-cutting; ‘G’ – gap-cutting; ‘P’ – preparation cutting
and ‘R’ – retention tree group. Letter ‘d’ in the variable abbreviations refers to the differences
from the mean values measured in the ‘Control’ plots. Full circles show the mean; vertical
lines denote the standard deviation of the samples. Letters designate the significant
differences among treatments and time (pairwise comparisons based on the linear mixed
models: Tukey-test, alpha=0.05). The horizontal blue line shows the level of control.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Spring Autumn
a a a a a a a a
Treatment
CC G P R CC G P R
(mean±
SD
)d S
oil p
H
a b
-0.2
0.0
0.2
0.4
0.6
Spring Autumn Spring Autumn
a a a a a ab ab b
Treatment
CC G P R
Treatment
CC G P R CC G P R
(mean±
SD
)d L
itte
r pH
(mean±
SD
; g/m
2)
d L
itte
r m
ass
a a a a a a b b
a b a b
CC G P R
-500
0500
1000
0.0
0.5
1.0
Spring Autumn
a a a a
Treatment
(mean±
SD
; %
)d L
itte
r m
ois
ture
conte
nt
a a a b
a b
CC G P R CC G P R
-10
010
20
a) b)
c) d)
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
Figure 5(on next page)
Diurnal pattern of selected microclimate variables.
Diurnal magnitudes of PAR (photosynthetically active radiation), VPD (vapor pressure deficit)
and Tsoil (soil temperature) among treatments in the peak of the growing season (i.e. June,
July, August; left) and during the transition period (March, April, September, October; right),
respectively. Lines represent means calculated by LOWESS function (based on the 6
replications for each variable per month), bands are 95% confidence intervals. Colors are
coding the treatments: control – green; clear-cutting – red; gap-cutting – purple; preparation
cutting – orange and retention tree group – blue. Note that scales of y-axes vary among the
graphs.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018
PA
R (
µE
m-2
s-1
)V
PD
(kP
a)
1600
1400
1200
1000
800
600
400
200
0
1.75
1.60
1.45
1.30
1.15
1.00
0.85
0.70
0.55
0.40
0.25
Time (hh:mm)
Tsoil
(°C
)
27
26
25
24
23
22
21
20
19
18
17
16
31
30
29
28
15
PA
R (
µE
m-2
s-1
)V
PD
(kP
a)
Tsoil
(°C
)
550
500
450
400
350
300
250
200
150
100
50
0
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.0512
11
10
9
8
7
6
5
4
Time (hh:mm)
00:00 06:00 12:00 18:00 24:00
00:00 06:00 12:00 18:00 24:00
Summer Transition period
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.26643v1 | CC BY 4.0 Open Access | rec: 8 Mar 2018, publ: 8 Mar 2018