Introduction

Changes in temperature and precipitation patterns predicted under future scenarios of global warming will have profound effects on primary productivity, plant species diversity and ecosystem function (Wang et al. 2011; Rózsa and Novák 2011; Williams et al. 2012; Vicente-Serrano et al. 2015). All of these factors will feed back to alter the quality and quantity of detrital inputs to soils, further altering patterns of nutrient cycling (Biró et al. 2012; Tóth et al. 2013) and soil organic matter (SOM) content and dynamics. Indirect changes in soil temperature and moisture regimes from global climate change will also affect the processing of litter by microbes (Chapin et al. 2009; Bond-Lamberty and Thomson 2010; Hagedorn et al. 2010; Wang et al. 2014). While both of these factors – litter production, and soil warming and/or drying - are included in most models of soil carbon balance, indirect effects of litter production and surface litter depth on soil microclimate, and thus microbial processing of litter and SOM, are less well studied. Detritus thus plays two major roles in terrestrial ecosystems: the above- and belowground litter effect on SOM, and litter forms a layer on the soil surface that affects microclimate (Sayer 2006).

The aim of this work was to explore effects of changing detrital inputs on soil microclimate in a Quercetum petraeae-cerris community in northeast Hungary. The Síkfőkút DIRT experiment constitutes an important part of a long-term international project that involves five experimental sites in the USA (Andrews Experimental Forest, Bousson Experimental Forest, Harvard Forest, University of Michigan Biological Station, Santa Rita Experimental Range) one in China (Huitong Natural Research Station of Forest Ecosystems) (Wang et al. 2013) and one in Germany (Universität Bayreuth BITÖK). The overall objective of the DIRT project is to explore how changes in the quality and quantity of detrital inputs affect soil physical, chemical and biological parameters (Nadelhoffer et al. 2004; Lajtha et al. 2005).

Several studies have already examined the effects of global warming on our experimental site. Longer and more severe drought periods are expected in the near future for several Central European ecosystems. Long-term meteorological data have clearly indicated that the climate of the forest has become drier and warmer over the past few decades with annual precipitation decreasing by 15–20 % in many Hungarian territories (Antal et al. 1997; Galos et al. 2009). The summer climate of the Carpathian Basin has shifted towards a more Mediterranean like climate (Domonkos 2003; Bartholy et al. 2007). Species composition and structure of the Síkfőkút forest have changed since the early 1970’s (Kotroczó et al. 2007; Tóth et al. 2007), with a 68 % decline in sessile oak (Quercus petraea), 16 % decline in Turkey oak (Quercus cerris), and the share of field maple (Acer campestre) has increased from 0 to 28 %. This has resulted in changes in detritus amount and composition. Mean leaf-litter production has declined from 4060 kg ha−1 y−1 between 1972 and 1976–3540 kg ha−1 y−1 between 2003 and 2010 (Kotroczó et al. 2012).

We hypothesized that detrital litter layer thickness would significantly affect annual, seasonal and daily temperature fluctuations of surface mineral soils by acting as a buffer that would also affect soil moisture levels. We also hypothesized that differences in soil moisture content among treatments would affect diurnal soil temperature range (DTR), with soils from plots without roots having greater soil moisture content and thus greater DTR.

Material and methods

Site description

We carried out our research in the Síkfőkút Experimental Forest in northeastern Hungary. The project area (27 ha) is located in the southern part of the Bükk Mountains at an altitude 325–345 m (47°55′N; 20°26′E). The area has been protected and is part of the Bükk National Park since 1976. According to Antal et al. (1997) the mean annual temperature (MAT) is 10 °C and the mean annual precipitation (MAP) is 553 mm. July is the hottest month with an average temperature of 20.2 °C and January is the coldest with −1.2 °C (Antal et al. 1997). This forest is a semi-natural stand (Quercetum petraeae-cerris community) with no active management since 1976 (Jakucs 1985). In this previously coppiced forest, the Sessile oak and Turkey oak species that make up the overstory are a hundred years old. Based on the data from 2003 to 2006, litter production consists of the following tree species in decreasing order: Sessile oak, Turkey oak, Field maple, and Cornelian cherry (Cornus mas). During the same period the average dry leaf-litter production was 3585 kg ha−1 and the average amount of total aboveground dry detritus (including branches, twigs, fruit and buds) was 6230 kg ha−1 (Tóth et al. 2007). According to the FAO World Reference Base, the soils are Luvisols (Świtoniak et al. 2014) with a pHH2O in surface soils (0–15 cm) without detrital manipulation of 5.2, and with a SOM of 39 g kg−1 dry soil (Tóth et al. 2013; Fekete et al. 2014).

The experimental detrital manipulation plots were established in the Síkfőkút DIRT site in November 2000. We established six treatments, each with three 7 × 7 m replicate plots (Fekete et al. 2011). There was a Control (CO) treatment. We applied two litter addition treatments. In Double Litter (DL) plots the amount of leaf litter was doubled. In Double Wood (DW) plots the amount of wood detritus (branches, twigs and bark) was doubled, and that of litter detritus was normal. In No Litter (NL) plots the aboveground detritus, in No Roots (NR) plots the living roots, and in No Inputs (NI) plots both the aboveground detritus and living roots were totally removed (Fekete et al. 2012). The project is belong to the international DIRT sites, its data are presented on its site: http:/cropandsoil.oregonstate.edu/content/data-inventory. The surface solar radiation was approximately the same in all treatments, as the distribution and slope of land did not show great differences (average slope of 5°) and the site faced south, so the climatic effect was similar in all plots. Moreover, the plots were established at random, thus reducing the effects of incidental minor differences.

Soil temperature was measured with ONSET StowAway TidbiT type data loggers (Onset Computer Corporation, USA) placed into the middle of each plot at 10 cm depth. Air temperature was measured 0.5 m above the ground with the same type of data loggers. Data loggers were programmed to measure soil and air temperature every hour from 06.17.2004 to 06.16.2008. The temperature data were grouped into seasons (winter: December, January, February; spring: March, April, May; summer: June, July, August; autumn: September, October, November). Soil moisture content at upper 12 cm depth was determined with a FieldScout TDR 300 (Spectrum Technologies Inc., USA) in all plots, every month.

Statistical analyses

Statistical analyses were performed using Statistica 7.0. Random sampling and the independence of samples were ensured by the experimental design. Experimental data were statistically evaluated by one-way ANOVA (assumptions were tested by Levene’s test for homogeneity of variances and Chi-square test for normality), linear regression and one-way ANCOVA. When groups were significantly different, ANOVAs were followed with Tukey’s HSD test. We analyzed the effects of air temperature on soil temperature in the treatments by linear regression, and differences among slopes were tested using one-way ANCOVA. When “p” ≤ 0.05, examined values were considered to be significantly different.

Results

The effects of detritus treatments on soil temperature and soil moisture content

Detritus treatments significantly influenced soil microclimate. The mean annual temperature values did not show any significant differences between the treatments (Table 1). However, the differences were significant within the winter (F(6;2191) = 70.31; p < 0.01) and the summer periods (F(6;2583) = 65.2; p < 0.01) (Table 2), but there were no significant differences for the transitional seasons (autumn and spring). In summer, air temperature was significantly higher than in all soil treatments. Soil temperatures in the root exclusion treatments (NR, NI) were significantly higher than in the any other soil treatments, and NL was significantly higher than in CO and the litter addition treatments. In contrast, air temperature was significantly lower than all soil treatments in winter. The temperatures in the aboveground litter exclusion treatments (NL and NI) were significantly lower than the CO, and litter addition treatments in winter. Similarly, the number of frost days, when the soil temperature was below 0 °C, was significantly different among treatments (Table 1). Therefore, the annual fluctuation of soil temperature was much higher in the detritus exclusion treatments than in CO or litter addition treatments (Figs. 1 and 2), and air temperature showed higher fluctuations seasonally than did soil temperature in the different treatments (Tables 1 and 2).

Table 1 Soil and air properties in the Síkfőkút DIRT treatments
Full size table
Table 2 Air and soil temperature (°C) and moisture (v/v%) in the Síkfőkút DIRT treatments during summer periods; (June, July, August) and winter periods (December, January, February) from 06.17.2004 to 06.16.2008
Full size table
Fig. 1
figure 1

Temporal variation of the summer soil temperature for Double Litter, Control and No Imput treatments

Full size image
Fig. 2
figure 2

Temporal variation of the winter soil temperature for Double Litter, Control and No Imput treatments

Full size image

The relationship between air temperature and soil temperature in the six treatments was shown by regression analyses. Air temperature and soil temperatures were significantly related when analyzed within each season based on daily averages (Table 3). Soil daily mean temperatures (DMT) in the root exclusion treatments (NR, NI) were most responsive to air DMT in summer (Fig. 3), and soil DMT in surface litter exclusion treatments (NL, NI) were most responsive to air DMT in winter (Fig. 4). The correlation was stronger in summer than in winter for all treatments. The regression analysis between air DMT and soil DMT exhibited significantly higher slope values in the regression equation in case of the NR and NI soil of the treatments than in the CO and litter addition plots in summer. However, there were no significant differences in slope values for the winter periods. The homogeneity of slopes between season values showed significant different among detritus exclusion treatments. These treatments showed significantly higher values in summer than winter (Table 3).

Table 3 The relationship between the soil temperature (°C) and air temperature of the different treatments in the summer and winter periods from 2004 to 2008
Full size table
Fig. 3
figure 3

Summer hourly temperature profile of the soil temperature in the different treatments and the profile of air temperature between 06. 01. 2005 and 09. 30. 2005

Full size image
Fig. 4
figure 4

Winter hourly temperature profile of the soil temperature in the different treatments and the profile of air temperature between 01. 01. 2008 and 02. 29. 2008

Full size image

Soil moisture contents were significantly higher in root exclusion treatments (NR and NI) than in the other treatments (F(5240) = 18.21; p < 0.001) (Table 2). There were no other differences in soil moisture among the other treatments (Table 2). Regarding moisture content values, the coefficient of variation were the lowest in root exclusion treatments, which showed that soil moisture content varied the least in these soils.

The effects of detritus on DTR

Hourly soil and air temperature readings were used to determine minimum and maximum daily temperatures, and the differences among minimum and maximum temperatures were used to calculate DTR. DTR of air was always significantly higher than DTR of the soils in all treatments (F(6;1526) = 812.8; p ˂0.001). The DIRT treatments significantly affected soil DTR fluctuations (Table 4). As temperatures in winter fell to below 0 °C, rapid temperature decreases were first seen in plots not covered by litter, followed by CO and DW plots; soil temperatures below freezing were not observed in DL plots. As temperatures rose above freezing, frozen soils had a 2–3 day lag due to isolating ability of the frozen upper soil layer, in winter, there were no significant differences in soil DTR among plots. In spring and summer, soil DTR was significantly smaller in plots with a litter layer (DL, DW and CO) than in detritus removal plots (NL, NR, and NI). In autumn, DTR was significantly smaller in DL, DW and CO than in NR and NI (Table 4).

Table 4 Air and soil values of DTR (°C) in the Síkfőkút site based on the means of all treatments from 17. 03. 2004 to 29. 02. 2008
Full size table

Discussion

Changes in soil temperature and moisture content

Various detritus treatments and soil biological processes both directly and indirectly influence soil microclimate (Sayer 2006). Therefore, the significant differences in temperature among the soils of the treatments during the winter and summer periods are considered to be the consequences of detritus treatments. In winter (especially when there is no snow) the thickness of the detritus layer had a profound effect on soil temperature. This insulating effect was observed in DL treatments, as temperatures in DL soils never fell below 0 °C. Soil temperature may also be influenced by soil biochemical processes (Raich and Tufekcioglu 2000; Bernhardt et al. 2005); decomposition of organic matter and other microbial processes can release a lesser amount of heat (Khvorostyanov et al. 2008). In the colder periods exothermic decomposition processes were significantly greater in litter addition treatments and CO than in detritus removal treatments, as shown by soil respiration values (data in Fekete et al. 2014; Kotroczó et al. 2014). These higher respiration rates could both be partially due to higher temperatures in litter addition plots due to insulation effects, and could also contribute to warmer conditions. In contrast, in summer NR plots had the highest soil respiration values among the treatments and had the highest soil temperature and moisture content as well. Because the NR treatment had no labile root inputs and no root turnover, the higher respiration rates were clearly due to microclimate effects.

NL and NI soils had lower albedo values than the lighter color surface covered by dry detritus, so they certainly absorbed more heat in summer, when solar radiation was intense (Bonan 2008). This was especially true for NI plots, whose soil was much darker due to its higher moisture content. The soil moisture values also showed large differences among the detritus treatments (Veres et al. 2013, 2015). Due to the lack of transpiration, the soils in the root exclusion treatments (NR and NI) were significantly wetter than soils in the other treatments, and they stayed moist even under severe summer drought (Fekete et al. 2012). A similar trend was found in the American DIRT sites. The soil moisture content of NR had higher with 86 % than the moisture content of CO in Síkfőkút DIRT site, while this difference was 9.3 % in Andrews DIRT site in Oregon (USA) and 17.5 % in Bousson DIRT site in Pennsylvania (USA) (Brant et al. 2006). These differences between the American and Hungarian DIRT sites may be explained by climate factors. Annual precipitation at Síkfőkút is much lower than that at the US DIRT sites (Andrews: 2370 mm yr−1; Bousson: 1050 mm yr−1), and rainfall is more seasonally distributed (Sulzman et al. 2005; Crow et al. 2009). Moreover, the annual mean temperature at Síkfőkút is higher than that at the US DIRT sites, so here’s higher evaporation.

Surface detrital layers also regulate soil water content (Ogée and Brunet 2002) by reducing evaporation from mineral soil while absorbing a fraction of the precipitation. Soil moisture content may also influence soil thermal conductivity; the higher the soil moisture content, the higher its thermal conductivity (Blackburn et al. 1998; O’Donnell et al. 2009). In contrast, the heat of evaporation cools the soil surface, and this effect is greater in moister soils. The differences among the DTR in the soils of the treatments likely reflected this evaporation effect. The litter layer of the NR plots is similar to CO plots, but the DTR is significantly higher in NR than in CO.

In addition, the absence of trees and shrubs in NR and NI plots leads to a lack of a shading effect. Therefore, it was the two root exclusion treatments that showed the highest temperature values in summer and the lowest ones in winter – in this latter case along with NL. Many factors in addition to soil water content, such as organic matter content, also influence soil thermal conductivity (Blackburn et al. 1998; Al-Shammary and Al-Sadoon 2014). Soil mineral particles have much higher thermal conductivity than soil organic matter, e.g., quartz has 14 times higher conductivity than organic particles (Farouki 1986; Perry et al. 2011). Thus soils with higher organic matter content warm more slowly than those with lower organic matter and higher mineral matter contents – provided other parameters are the same. Soil organic carbon content was 51.9 g kg−1 dry soil in the upper 5 cm soil layer in CO. Soil carbon content was higher in DL, DW with 29.8 and 18.3 % and lower in NL, NR, NI with 18.2, 11.6, 19.5 %.

Clearly detrital thickness can reduce the effects of soil temperature extremes and moderate minimum and maximum temperature values, creating a more balanced microclimatic for soil organisms. Increase in soil moisture content, which is in connection with litter layer thickness and with lack of living roots, has significant effect on soil temperature. Our results show that average annual dry leaf litter amount of 0.354 kg m−2 moderates the soil cooling off in winter by 1.1 °C and soil warming in summer with 0.6 °C. It also moderates DTR by 0.6 °C, and annual maximum temperature fluctuation by 3.8 °C. In root exclusion treatments annual maximum temperature fluctuation by 3.5 °C, DTR with 0.5 °C was higher due to lack of living roots than in treatments containing living roots. If the climate becomes warmer and drier litter production may decrease, creating a thinner litter cover, which may increase daily and seasonal temperature extremes. The effects of litter layer thickness on soil processes is important to include in earth system models that aim to predict soil carbon stocks and soil respiration.