Effects of dissolved char on microbial metabolism in forest soils

Effects of dissolved char on microbial metabolism in

forest soils

MSc thesis Earth Sciences

Spruce forest, April '19

Evy de Nijs

University of Amsterdam

Master Earth Sciences

Geo-Ecological Dynamics Track

Amsterdam, January 27

th

_{, 2020}

Supervisor: dr. A. Tietema

Assessor: dr. J.R. Parsons and dr. ir.

W.E. Morriën

1

Abstract

Climate change is associated with an increased risk in the prevalence of wildfires. Forests store large amounts of C which are threatened by these wildfires. Pyrogenic material remaining on the forest floor after a wildfire can either be mineralized or sequestrated into the soil. Char constitutes an important part of the stable SOC pool in forest soils. This study investigated the effects on microbial metabolism in forest soils upon the addition of wood-derived dissolved char from different species and at different stages of decay. Char was produced under laboratory fire conditions and extracted in water after which microbial activity was measured for a 4-week period in a liquid incubation. Further stages of decay resulted in increased flammability with higher peak temperatures and combustion completeness. For the beech samples, it furthermore resulted in a decrease of extractable C but a higher proportion of stable C. The CUE of the microbial communities increased during incubation, indicating adaptation towards using char as energy source, with no trends in differences between species or decay stage. This implies that C from pyrogenic material could be sequestrated in the soil in the form of microbial C. Thereby it partly mitigates the rapid C release during a forest fire by enhancing subsequent microbial C storage in soils. This study provided a promising workflow to assess the effects of dissolved char on microbial metabolism by mimicking natural fire conditions. It also indicated the need of future research to further elucidate the underlying mechanisms. Keywords: forest fire, coarse woody debris, decay, dissolved char, microbial biomass, respiration, CUE, metabolism Reference style: APA 6th _edition

2

Contents

Abstract 1 Introduction 3 Materials and methods 5 Branch sampling 5 Wood properties 5 Production of Dissolved Char 6 Characterization of DC 7 Incubation experiment 7 Measuring microbial activity 7 Data analysis 8 Results 9 Wood characteristics 9 Fire experiment 10 Dissolved char characterization 11 Incubation DC: microbial activity 12 Discussion 14 Combustion of course woody debris 14 Characteristics of dissolved char 14 Microbial metabolism as a result of dissolved char 15 Synthesis 16 Conclusion 17 Acknowledgements 18 References 19 Appendix 21 a. FTIR peak identification 21 b. Incubation graphs: respiration, microbial biomass & CUE index ratio 22

3

Introduction

An increase in atmospheric concentrations of greenhouse gases causes our climate to change (IPCC, 2013). Forests ecosystems are effective in mitigating climate change by acting as a sink for carbon (C), component of the greenhouse gas carbon dioxide (CO2). Forests capture atmospheric C and

subsequently sequester it in biomass (Kilpeläinen et al., 2016) and in soil organic matter (SOM). The latest IPCC report predicts that global warming will cause alterations in the global water cycle which results in an increased prevalence of extreme weather events (IPCC, 2013; Orth et al., 2016). Consequently, increased occurrence of extended periods of severe drought can be expected, which are associated with an increased risk of wildfires (Diffenbaugh et al., 2015). Wood biomass is an important pool of organic C, hence understanding the fate of this C is crucial in predicting C fluxes. Dead woody biomass accumulates on the forest floor and the released C is thought to be sequestrated in the soil, thereby enhancing the sink strength of forest ecosystems (Magnússon et al., 2016). Decomposition of woody biomass may take up to several years under standard conditions, during which C can dissolve, enter the soil and eventually be transferred to the stable soil organic C (SOC) pool. The fate of wood C in relation to wildfires is twofold. During a fire, the stored C partly returns to the atmosphere via rapid release. After a fire, the residues laying on the forest floor might further be decomposed, incorporated directly into the soil or the C might dissolve and enter the soil similar to dead woody biomass (Cornwell et al., 2009; Zhao et al., 2018). The balance between these processes (rapid release or slow incorporation) is critical when assessing the extent to which forests act as a C sink. The effects of forest fires on C fluxes is imperative in the context of climate change since forests can either enhance climate change by positive feedback or mitigate climate change via negative feedback. Fire under normal conditions results in incomplete combustion of biomass which produces the C-rich residue called pyrogenic organic matter (PyOM), also been referred to as ‘biochar’ or ‘black C’. The C-rich PyOM stored in soils is part of the total soil C pool, which is the greatest pool of terrestrial C and stores more C than vegetation and the atmosphere combined (Houghton, 2007). Because the SOC pool is relatively large, even slight changes in C-fluxes will have significant impact on the global C balance (Stockmann et al., 2013). PyOM-derived C is considered a significant component of the SOC pool in forest soils. About 1% of the above-ground biomass is estimated to return to the soil in various forms of char, with wildfires as the largest contributing source globally (Sohi et al., 2010). Studying the extent and mechanisms underlying the incorporation of char-derived C in soils is essential to enhance our understanding on the behaviour of char in soils and hence assess the effects of forest fires. (Masiello, 2004; Cheng et al., 2008; Sohi et al., 2010; Lian & Xing, 2017).

Incomplete combustion chemically alters wood resulting in a complex mixture of C-rich combustion products. The residence time of soil incorporated char varies strongly by degree of compositional complexity of the by combustion produced char (Saiz et al., 2015). Different theories regarding the composition and complexity of this material exist. One concept represents PyOM as a continuum of C-rich products ranging from slightly charred biomass to soot (Masiello, 2004). Another widespread concept assumes that PyOM is a heterogeneous mixture consisting of relatively small aromatic biopolymers with high levels of substitution allowing for rapid oxidation (Knicker, 2007). Decomposition changes both the structure and chemical composition of wood and thereby strongly influence how it will be consumed by fire (Zhao et al., 2014). There is full agreement that the structure of PyOM depends on both the origin of the biomass and conditions during combustion (Masiello, 2004; Sohi et al., 2010; Bantle et al., 2014; Lian & Xing, 2017). Decay stage strongly influences combustion conditions and thereby regulates the composition of char. One study showed that with increasing decay stage, and associated decreasing wood density, the flammability as well as gas emissions increased significantly during a fire event (Zhao et al., 2018). Flammability is an umbrella term describing the fire properties of a material. It is defined as ignition ease, duration, temperature intensity and the total amount of combusted fuel (Lowden & Hull, 2013; Zhao et al., 2018). Tree species together with moisture content also influenced the combustion process and products, but to a lesser extent. During combustion, peak process temperature, heating length and rate have the most significant influence on the resulting residue (Sohi et al., 2010; Lowden & Hull, 2013). Slow and low temperature pyrolysis, meaning gradual temperature increase until approximately 450-550 °C in the total absence of oxygen were found to be optimal for the production of char (Lowden & Hull, 2013). Most studies on char characteristics are performed under laboratory-controlled circumstances with limited to no oxygen availability where

4

different predetermined temperatures and times are maintained to assess its effect. However, wildfires do not take place under controlled conditions and therefore combustion conditions depend on the specific field circumstances. During a wildfire only a part of the combustion will occur under limited oxygen availability resulting in the production of char, which is incompletely combusted biomass (Sohi et al., 2010). The solubilisation and fate of wildfire originated char located on forest floors is still an unknown pathway. Dissolution of char by precipitation results in dissolved char (DC) leaching into the soil matrix. This DC might be lost via leaching to deeper soil layers and groundwater or be sequestrated in the soil as stable SOC with the help of soil microorganisms (Knicker, 2007; Cotrufo et al., 2013). It is proposed that the presence of dead woody biomass on forest floors might enhance the abundance and diversity of microbial communities which play an important role in the belowground stabilization of dead wood derived C (Magnússon et al., 2016). This is also supported for the addition of char by a study focussing on the effects of biochar on microbial communities. During a long-term incubation experiment, the addition of char to soil resulted in increased carbon use efficiency (CUE) of the microbial community and changes in its composition (Jiang et al., 2016).

The stabilization of SOM is mainly governed via limitation of C accessibility to decomposers through both physical and chemical protection mechanisms (Dungait et al., 2012; de Nijs & Cammeraat, 2020). Physical protection takes place through the occlusion of C in soil aggregates and chemical protection through organo-mineral associations (Schmidt et al., 2011; Berhe & Kleber, 2013). Cotrufo and colleagues (2013) proposed the ‘Microbial Efficiency-Matrix Stabilization’ (MEMS) framework in which they emphasize the role of microbial derived products of decomposition in promoting aggregation and the formation of organo-mineral associations. The input of easy accessible (‘labile’) C for microbes results in higher microbial CUE and hence more microbial C and microbial-derived compounds, potentially leading to increased levels of stable SOC formation. This concept is supported by a study showing that deprivation of fresh SOM input for an extended period of time leads to a drastic decline of stable SOM levels (Hirsch et al., 2009; Regelink et al., 2015).

DC provides soil microorganisms with fresh C, however, the degree of complexity and hence accessibility of this C supply is important with regard to its fate. Hilscher et al. (2009) showed that the microbial degradation of PyOM can start rapidly after a fire event, thereby supporting the heterogeneous mixture concept of Knicker (2007). This suggest that it might not be just the complexity of char molecules, known as recalcitrance, leading to the long-term presence of char in soils but rather processes associated with the diversity of C-rich molecules. The ‘labile’ C molecules could enhance microbial activity and hence microbial derived compounds which promote the formation of stable SOC from the remaining char compounds according to the MEMS framework (Cotrufo et al., 2013). This process could be supported in a soil system by the observation of relatively low respiration after DC addition to soil in combination with substantial microbial biomass, hence a high CUE. This has previously been shown for the addition of solid biochar to soils in a 30-month laboratory incubation by Jiang et al. (2016). If the addition of DC indeed leads to an increase in CUE, associated with enhanced C sequestration, the interaction between DC and microbial communities could be an underlying mechanism mitigating the rapid C releases associated with forest fires. The underlying mechanisms of the influential role of soil microbial communities on char leachates have so far remained elusive. An insight in the effects of DC on the functioning of microbial communities in forest soils could elucidate the mechanisms involved in the long-term microbial stabilization of DC. The aim of this research was to increase our understanding of the role of microorganisms in the fate of DC in order to optimise management strategies in regard to maintaining the C sink service of forest ecosystems. To mimic the effects of wildfires the intention was to remain as close to a real field situation as possible under laboratory conditions. This study focused on the effects of DC derived from PyOM from different tree species at different degrees of decay on the activity and efficiency of microbial communities in forest soils. The following research question was addressed:

‘What mechanism underlies the interaction between wood-derived dissolved char and the functioning of microbial communities of forest soils in respect to the fate of carbon?’

5 This research question is two-fold and integrates the fields of biochemistry and microbial ecology, therefore two sub-questions were formulated: 1) What is and what regulates the composition of dissolved char? 2) What is the influence of dissolved char from different tree species and stages of wood decay on the functioning of soil microbial communities? It was hypothesised that (1) the composition of DC is a heterogeneous mixture of C-rich water-soluble compounds which is governed by the tree species and stage of decay of the wood. Later stages of decay were expected to result in more complex DC compositions (Hilscher et al., 2009; Zhao et al., 2018). Furthermore, it was hypothesized that (2) the addition of DC induces an increase in microbial substrate efficiency. This might subsequently enhance the formation of stable SOC in a soil system (Cotrufo et al., 2013; Jiang et al., 2016; Magnússon et al., 2016).

Materials and methods

Branch sampling

The wood samples were collected from a forestry plantation located in the Hollandse Hout forest in the Netherlands (52.4656° N, 5.4412° E). This site is located below sea-level and was reclaimed from the sea in the 1960-ies, hence a marine clay soil can be found which is rich in calcium carbonate (CaCO3). Sampling took place mid-April 2019 at the start of the growing season, after a week of no precipitation with an average daily temperature of 9 °C (KNMI). Wood samples were obtained from 2 tree species of different higher clades (e.g. gymnosperm and angiosperm). Spruce (Picea abies) was collected at plantation number 161T which was planted in 1972, beech (Fagus sylvatica) was collected from plantation number 130A which was planted in 1970. Dead wood branches were collected from the forest floor with a diameter of 5-9 cm (avg 6 ± 0.2 cm). Both external appearance and texture of the branches were used to identify the 3 decay stages: freshly dead, partly decomposed and very decomposed (Zhao et al., 2018). For each species 3 samples per decay class were sampled to represent the decay continuum, resulting in a total of 18 samples. Furthermore, mixed organic litter from the top 5 cm was collected for both sampling locations. The branches were sawed into 3 pieces, the main sample of 15 cm and 2 subsample discs of 2.5 cm from each side of the main mid-sample. The ends of the branch were discarded to ensure representative (sub-)samples. All samples were stored airtight at 4 °C in darkness until further analysis.

Wood properties

Wood density was used as indicator of stage of decay and measured according the water-displacement method using one disc. The water-water-displacement method calculates the basic specific gravity of wood as the dry biomass per unit volume of green wood (Williamson & Wiemann, 2010). Each disc was weighed for initial moisture content calculation (green mc). Since green volume measurements are influenced by the porosity and moisture content of the sample the discs where submerged for 48 hours to ensure complete saturation. Volume was measured according to Archimedes’ principle, where a beaker with demineralized water was placed on a balance and zeroed, the saturated disc was added and submerged using 4 volumeless needles (Fig. 1). The water weight of the displaced volume is equal to the sample volume (Chave, 2005). Oven-dry mass was obtained by drying the saturated samples at 105 ˚C for 120 hours to remove all free and bound water (Williamson & Wiemann, 2010) (WTC binder FED-240). Wood density was calculated as the ratio of dry oven-dry mass divided by its saturated volume.

The other discs were air-dried in an incubator (Termaks) under constant ventilation at 30 °C for 22 days until no further water weight loss was observed. The air-dry subsamples were used to assess C content and C/N ratio on the elemental analyzer (Vario El cube) using sawdust collected in triplicate from each disc (Fig. 2). Finally, pre-burn moisture content of the air-dry samples was determined (gravimetric, 93 hr at 105 °C).

6 Figure 1 (left) Volume measurement according to the water-displacement method. Figure 2 (right) Sawdust collection of the air-dry subsamples for C & N analyses.

Production of Dissolved Char

The main branch samples were air-dried in an incubator (Termaks) under constant ventilation at 30 °C, to avoid loss of volatiles, until no further water weight loss was observed for 27 days. The air-dry branches had a mean moisture content of 6.0% ± 0.1. All fires were conducted in a metal basket of 45 by 25 cm under a fume hood via ignition of a fuel bed under normal atmospheric conditions. The laboratory was under constant ventilation and with an average temperature of 23 °C. To mimic the spreading of a natural fire a fuel bed was built which acted as ignition source for the branches. The fuel bed consisted of 50 g wood wool (mc 8.8% ± 0.3 n=3) with on top 200 g of 10 mm beech wood chips (mc 7.9% ± 0.01, n=3). Each branch was placed in the middle of the ignition bed and fire was ignited from two sides using two cotton pads injected with 1 ml of 96% ethanol (Fig. 3). Figure 3 Set up of the burning experiment. A) Ignition bed with the branch located on top and 4 thermocouples positioned above the branch B) Fire was ignited with 2 ethanol-injected cotton pads at each side of the metal basket. Combustion temperature was recorded every second using 4 thermocouples (1 mm thick type K thermocouple, TC Direct, Uxbridge, UK), which were placed evenly above the branch at 2 cm distance (Fig. 3). This was used to estimate the average maximum reached temperature of the fire. The temperatures of 6 isolated fuel bed fires were recorded to calculate the baseline temperature of the fuel bed. Furthermore, time until the fire reached the branch, ignition time defined as first flames reaching the branch until the first flame coming out of the branch and the duration of the fire defined as flaming was monitored (Zhao et al., 2018).

15 minutes after ignition, the smouldering branch was removed from the fuel bed to further smoulder on its own without the influence of the ignition material. In most cases flame extinction of the branch had occurred by this time, otherwise the flaming stopped within 1 minute after removal. Approximate combustion completeness was calculated as percentage biomass loss during burning. The PyOM was stored air-tight at 22 °C from mid-May until mid-September.

7

Char was scraped from the burned samples, mortared and extracted with ultrapure water at a 1:80 char to water weight-ratio for 24 hr under continuous shaking. The extracts were centrifuged at 4000 rpm for 15 minutes and the supernatant filtered through a 0.45 µm membrane. Extracts were refrigerated at 4 °C in darkness until further analysis.

Characterization of DC

All compositional characterization was conducted using the 1:80 char to water extracts. TOC represents DOC since particulate organic C (POC) was filtered out. DOC and TN content of the dissolved char extracts was measured using the Shimadzu TOC elemental analyzer (Shimadzu TOCVCPH with

additional TNM-1 unit).

Specific ultraviolet absorbance 254 (SUVA254) is the absorbance at wavelength 254 nm normalized

for the DOC concentration (Eq. 1) and used as indicator for DOC aromaticity and hence molecular complexity (Weishaar et al., 2001; Jamieson et al., 2014). 𝑆𝑈𝑉𝐴%&'= 𝐴%&'/[𝐷𝑂𝐶] (1) Spectral slope (S275-295) was determined by fitting an exponential decay curve to the absorbance coefficients within the spectrum 275 to 295 nm (Jamieson et al., 2014). An increase in spectral slope indicates an increase in molecular weight (Helms et al., 2008). UV absorbance measurements were conducted on the Spectroquant Prove 300 photo spectrometer, using quartz cuvettes for the spectrum of 200-600 nm, at 1 nm increments. Fourier-transform infrared spectroscopy (FTIR) spectra were measured to indicate the presence of known functional groups indicated by absorption at specific wavelengths (Bantle et al., 2014). Samples were freeze-dried to exclude the effects of water and the extracted OM were measured for the spectra from 4000 to 400 cm-1 _{by accumulating 24 scans on the FTIR spectrometer (Bruker, OPUS).}

Incubation experiment

The prepared char extracts where inoculated and incubated (20 °C, dark room) for a total of 4 weeks to monitor the microbial activity induced by the availability of char as C source. 10 measurement points for respiration and microbial biomass were taken, starting with a 1-day interval increasing to a 7-day interval. Four very decomposed samples were not included in the incubation due to almost complete combustion resulting in an insufficient amount of char. The char extracts of the remaining 14 samples where diluted to a concentration of 20.0 ±1.0 mg C/L. To ensure a representative microbial community, an inoculum extracted from forest floor litter was added to each char extract. The inoculum was made by adding 70.0 g mixed litter from the spruce and beech forest to 1000 ml ultrapure (MilliQ) water and shaken for 60 min. The mixture was incubated at 20 °C for 8 days where subsequently the supernatant was filtered through 4-7 µm membrane. The inoculum was diluted to a concentration of 20.0 mg C/L, equal to the C concentration of the char extracts. For each sample and measuring point a 120 ml incubation bottle was prepared with 40.0 ml char extract and 10.0 ml inoculum. After closing, the headspace was flushed 4 times with synthetic air to an overpressure of 200 hPa. At each measuring point first a gas sample was taken followed by the microbial biomass analyses; measurement protocols are described in the next paragraph.

Measuring microbial activity

At each sampling point both microbial respiration and microbial biomass were determined. For microbial respiration, a gas sample of 10.0 ml from the headspace of the incubation bottle was taken and injected in a 20 ml vial which was filled with Helium and brought to a pressure of -200 hPa. Pre- and post-pressure of both the incubation bottle and vial was measured for calibration purposes. The CO2

concentration in the vial headspace was analyzed on a gas chromatograph (Trace GC Ultra, Thermo Scientific). Respiration values were calibrated and expressed as rate of µg respired C per mg substrate C per day.

Microbial biomass was determined following the direct chloroform extraction method, adjusted for complete liquid extraction (Gregorich et al., 1990; Setia et al., 2012). At each measuring point two times

8

20.0 ml for each of the incubated extracts was pipetted to two 50 ml centrifuge tubes (for the fumigated & non-fumigated samples). 500 µl ethanol-free chloroform was added to one set of the tubes, the fumigated samples. All tubes where shaken vigorously for 1-hour (150 rpm) whereafter the supernatants where left to settle for 10 minutes. Each sample was filtered through a Whatman 42 filter and filtrates where bubbled with Nitrogen for 30 minutes to remove any residual chloroform. TOC concentrations of each filtrate where determined the same day on the Shimadzu TOC elemental analyzer. The amount of microbial biomass was calculated as the difference between TOC concentrations of the fumigated and non-fumigated samples and expressed as µg microbial C per mg substrate C.

Data analysis

Relationships were tested on correlation by fitting linear models through the data, the R2_{-value was} used as measure of suitability where correlation was significant. Differences between the mean values of species were tested using a t-test. Time series data from the incubation experiment were analysed using the functional analysis method which focuses on the underlying pattern. Respiration data was fitted with exponential decay curves for each sample, each fitted parameter was modelled using a linear regression model with species and wood density as predictor variables and tested with an ANOVA. Microbial biomass and CUE-index showed a different pattern, so it was decided to transform up to a 4th

order polynomial and fit this with a linear regression model with species and density as predictor variables. Models with various response variables and interactions were evaluated and tested with ANOVA. Matlab R2019b was used for data visualization, curve fitting and statistical analysis. Significant differences were identified were p<0.05.

9

Results

Wood characteristics

Wood density was used as indicator of decay, where lower densities indicates further decay. The sampled branches covered the whole spectrum of decay, ranging from freshly dead via partly decomposed to very decomposed (Fig. 4). For spruce, the wood density ranged from 0.38 g/cm3 _{for the freshly dead branches to 0.18}

g/cm3 _{for the very decomposed branches. Wood density of the beech}

samples covered a broader range from 0.53 g/cm3 _{to 0.15 g/cm}3_.

There was no significant difference in mean wood density between species (p=0.09).

A positive relation between decay stage and green moisture content was present for both species (Fig. 5). Moisture content upon sampling was higher with increasing decay stage and thus decreasing wood density. This relationship was strongest for the spruce samples. However, the range of green mc was similar for both species with approximately 20% mc for the freshly dead samples and an approximate mc of 65% for the very decomposed samples.

C/N ratio did not show a significant relationship with decay stage for either species (Fig. 6). Average C/N for respectively spruce and beech was 166 ±22 and 129 ±14. There was no significant difference in mean C/N between the species (p=0.25). Independent C and N concentrations remained stable among decay stages for both species and no significant differences between species where found (p=0.28 & p=0.24). C and N concentrations for spruce were respectively 44.7 ±2.16 % and 0.32 ±0.04 % and for beech 47.0 ±0.33% and 0.42 ±0.06%. Figure 5 (left) Correlation between the green moisture content and wood density for spruce and beech (n=9 per species). Figure 6 (right) C/N boxplot for spruce and beech.

Figure 4 Wood density boxplot

for the sampled spruce and beech branches.

10

Fire experiment

Average flaming duration of spruce samples was 760 ±45 sec, while the beech samples flamed significantly shorter with an average duration of 555 ±27 sec (Fig. 7, p=0.01). No correlation between fire duration and wood density within species was found. Maximum reached temperature showed a significant positive relationship with decay stage for both species (Fig. 8, spruce p=0.03, beech p=0.02). Decreasing wood density resulted in higher maximum temperatures, with a steeper slope for spruce which can be attributed to the smaller wood density range in relation to decay stage. The maximum temperature reached by solely the fuelbed was 666 ±24 °C. The mean reached maximum temperature was significantly higher for beech samples compared to spruce samples (p=0.0326). Figure 7 (left) Fire duration boxplot for spruce and beech. Figure 8 (right) Correlation between mean maximum temperature and wood density for spruce and beech (n=9 per species). Fuelbed represents the maximum temperature for solely the fuelbed (n=6). Combustion completeness was strongly correlated to wood density for both species (Fig. 9, spruce p=0.01, beech p<0.01). Freshly dead samples only lost approximately 10% biomass whereas the very decomposed samples reached almost complete combustion.

Figure 9 Correlation between combustion completeness and wood density for spruce and beech (n=9 per

11

Dissolved char characterization

Dissolved char of both species showed a

negative relationship between DOC

concentration and wood density, however only for beech this trend was significant (Fig. 10, p=0.02). Overall, coarse woody debris in a further stage of decay resulted in a lower DOC concentration.

SUVA254 values were not significantly

correlated with wood density for the spruce samples (p=0.16), but there is a significant correlation for the beech samples (Fig. 11, p=0.04). For beech an increase in decay stage is correlated with increasing SUVA254 values

which indicate an increase in DOC aromaticity. No significant differences in SUVA254 values

between the species was present. All samples with DOC concentrations below 10 mg C/L where omitted from this analysis due to distorted outcomes. The S275-295 values for the spruce samples did not show a significant relationship with wood density. The beech samples did show a significant correlation with wood density, increased decay stage resulted in lower S275-295 values indicating an increase in molecular weight (Fig. 12, p=0.0426). No significant differences in S275-295 values between species were present. Figure 11 (left) Correlation between SUVA254 and wood density for spruce and beech (n=9 per species). Figure 12 (right) Correlation between S275-295 and wood density for spruce and beech (n=9 per species). FTIR analysis showed similar absorption spectra for the functional group ranges for both species (Fig. 13). No relation with wood density within species was observed. All samples showed a very broad but weak band in the range 3500-3000 cm-1_{, which can be attributed to O-H and N-H stretching}

(Whitman et al., 2013). Most samples showed weak absorbance for C=C vibrations and stretching in the range 1410-1670 cm-1. All samples showed weak to medium absorbance for C-H deformations in the range 1380-1430 cm-1_{, which can be attributed to phenolic OH and aliphatic CH groups (He et al., 2009).} Finally, weak absorbance in the C-O / C-O-C region between 1300 and 1000 cm-1 _{was observed for most} samples. Absorbance around 1140 cm-1 _{is characteristic for C-O vibrations, whereas absorbance around} 1000 cm-1 _{is characteristic for C-O-C vibrations (Norwood et al., 2013). The dissolved char is on the latter} Figure 10 Correlation between TOC concentration in the dissolved char extracts and wood density for spruce and beech (n=9 per species).

12 side of the spectrum which indicates the presence of the ring structure of carbohydrates. An overview table with the peak identifications can be found in appendix a. Figure 13 FTIR spectra of all spruce & beech samples. Most samples showed a weak absorbance for C=C vibrations and stretching (1410 – 1670 cm-1_{). All samples showed weak to medium absorbance for C-H deformation (1380 –} 1430 cm-1_{). Almost all samples showed a weak absorption in the C-O / C-O-C region (1300-1000 cm}-1_).

Incubation DC: microbial activity

Respiration rates showed a similar exponential decrease over time for both species, with little variation between the different decay stages (Fig. 14). Respiration rates per mg substrate started for both species at 16 ±1.5 µg C mg-1 _{C day}-1_{, declining to below 3 µg C mg}-1 _{C day}-1 _{after 21 days. Both species}

showed a relative increase in respiration between day 17 and 21, this phenomenon was slight stronger for beech. Exponential decay curves were fitted to the data and the resulting parameters were tested in a linear model. The y-intercept was not significantly different for either species, wood density or their interaction. The growth rate constant was significantly higher for beech, indicating a stronger decrease in respiration rate over time (p=0.005). Hence, the samples with spruce derived DC respired more C compared to those of beech derived DC samples during the incubation period of 28 days. Wood density within species had a significant effect on respiration rate (p=0.008), however with opposing effects for both species. Spruce showed a slightly weaker decrease in respiration rate with increasing decay stage while beech showed a stronger decrease constant with decay stage. Further stages of decay resulted, over time, in more cumulative respiration for spruce and less cumulative respiration for beech.

Microbial biomass showed fluctuating levels during the first days, followed by a relative stable period until day 18, after which biomass levels started to increase (Fig. 15). Beech showed relative higher biomass levels during the first 16 days compared to spruce. Hereafter, the spruce samples showed higher biomass levels until the end of the experiment. Microbial biomass increased in the first two days in the spruce samples, where the beech samples showed steady biomass levels, both decreased after this initial growth. Microbial biomass was predicted most accurately with a linear model including the interaction of species and time; biomass~day+day2_{*species. Decay stage had no significant effect on}

microbial biomass during the experiment. The interaction of species and time was significant with a lower growth rate for beech compared to spruce (p=0.003). Hence, after 28-days of incubation, the cumulative levels of microbial biomass of the samples with spruce derived DC were higher than those of beech derived DC.

13

Figure 14 (left) & 15 (right) Average respiration rates and microbial biomass normalized to resource

availability for incubated spruce and beech (n=7 per species).

Figure 16 Average CUE-index per species expressed as µg respired C per µg microbial C per day (n=7 per

species).

Determining absolute levels of CUE was not possible, however the relative shift in respiration to biomass yields an index of CUE expressed as µg respired C per µg microbial C per day (de Nijs et al., 2018). Data from the first 4 days showed large and unrepresentative fluctuations which can be attributed to the initial stabilization period, therefore this data was omitted from the analysis. The ratio of respiration: microbial biomass decreased over the course of the experiment indicating an increase in CUE. Beech showed a relatively lower CUE-index ratio, and hence higher CUE, compared to spruce in the beginning, which converged around day 19 (Fig. 16). The observed peak at this measuring point can be explained by the relatively higher measured respiration rates at this moment. Despite this peak in CUE-index around day 19, the overall trend was a significant decline in CUE-index, and hence indicates an increase in CUE of microbial communities when DC is added (p<0.005). Neither species nor decay stage had a significant effect on the ratio respiration: microbial biomass (p=0.31 & p=0.27). Respiration, microbial biomass and CUE-index graphs for the individual samples per species can be found in

14

Discussion

Combustion of course woody debris

Samples were taken from two different tree species at different stages of decay since this regulates dead wood properties and subsequent flammability. Wood density was used as indicator of decay stage, lower densities indicate further stages of decay. Initial density and subsequent density decline with further decay did differ between the species. Densities of the spruce branches fell in a narrower range compared to the beech branches. The field class ‘freshly dead’ for spruce started at a significantly lower density compared to the beech samples, and the decrease of wood density showed a slower decline (Weedon et al., 2009). This was expected since spruce belongs to the group of gymnosperms which often show lower living wood densities and slower decomposition rates in dead wood due to the higher lignin content compared to angiosperms to which beech belongs (Weedon et al., 2009; Zhao et al., 2018). Lignin content is directly linked to decay, decreasing wood density is associated with increasing lignin concentrations (Zhao et al., 2014). Measuring wood density accurately did prove difficult since wood is a porous material where pore volume increases with decay. The extent of wood saturation, and hence moisture content, upon starting the volume measurements is often omitted in literature. Along with different sampling strategies and drying temperatures this complicates the comparison of values found by other studies (Williamson & Wiemann, 2010). To ensure similar analysis for all decay stages, a 2-day submerge period was chosen for complete saturation combined with oven drying at 105 ˚C for all bound water to be removed (Williamson & Wiemann, 2010). Although disintegration of the samples during saturation or subsequent volume measurements was limited this might have had a minor influence on wood density. C/N ratio does not significantly influence the decomposition of course woody debris as long as it remains within the range of 40 to 400 (Freschet et al., 2012). For optimal microbial decomposer functioning a suitable C/N ratio is important. The C/N ratio of both tree species fell in the common range, with no significant trend with decay stage. Weedon et al. (2009) emphasized, in their meta-analysis, that the N concentration is an important controller of decay rate, since sub-optimal levels hinder microbial decomposition. The observed C/N fluctuations in this study can mainly be attributed to fluctuations in the measured N concentrations, these fluctuations where not correlated to decay stage. The average N concentration was slightly lower for spruce, though not significantly, which corresponds to the characteristics associated with gymnosperm wood (Weedon et al., 2009).

Green moisture content increased with increasing decay stage for both species, with a relative stable green mc of around 20% until a density of 0.3 g/cm3_{. Moisture content did not influence}

flammability in this study, since all samples were air-dried before burning. Flammability is moderately inhibited by moisture levels above 30% which are expected under natural fire conditions (Zhao et al., 2018).

Decay stage across tree species has previously been indicated as the predominant driver of wood flammability due to the increasing lignin concentration with decreasing density (Lowden & Hull, 2013; Zhao et al., 2018). This experimental laboratory fire measured the effects of species and decay on combustion according to the flammability parameters: fire duration, maximum reached temperature and combustion completeness (Lowden & Hull, 2013; Zhao et al., 2014). Spruce samples burned significantly longer than beech samples. But the longer fire duration and supposedly higher lignin concentration did not result in higher peak temperatures, the maximum reached temperature was highest for the beech samples. Decay stage did not prove to influence fire duration for either species. Decay stage did influence both maximum reached temperature and combustion completeness, which both showed a positive relation with decay stage. Overall, flammability increased with increasing decay stage for both species. A shorter fire duration with higher maximum reached temperatures indicated that beech is the more effective fuel source compared to spruce. It has also been shown that species, but foremost charring temperature, are the main drivers of char properties (Gundale & DeLuca, 2006).

Characteristics of dissolved char

Further stage of decay resulted in significant decreasing levels of DOC. Increased charring temperatures have been associated with higher extractable C content by Gundale and DeLuca (2006). Although the maximum reached temperature differed, no significant differences in DOC concentrations

15 of the DC between the species were observed. Furthermore, increasing decay correlated positively with increasing maximum reached temperature but the relation with DOC was opposite. A side note here is the temperatures used in the study of Gundale and deLuca (2006) laid far apart with 350 and 800 °C, whereas temperatures in this study ranged from approximately 500 to 720 °C. The negative relation between decay stage and DOC in DC is opposite to the relation found by Blonska et al. (2018), who looked at leachates from dead wood. It appears that charring affects the amount of extractable C stronger with increased decay stage and thereby shifts the relationship compared to dead wood leachates. They also demonstrated that the concentration of DOC in leachates was higher in the deciduous species compared to the coniferous species, which is not confirmed in this study (Błońska et al., 2018).

Decreasing DOC concentrations corresponded with increasing aromaticity and decreasing molecular weight for the beech samples, this pattern was not significant for spruce. Previous studies found that aromaticity was negatively correlated to the labile C fraction of DOC, suggesting an increase in recalcitrant aromatics with increasing decay stage for beech (Fellman et al., 2008; Jamieson et al., 2014). This could potentially result in a proportionally larger fraction of C from DC to be moved to the stable SOC pool. High aromaticity corresponded to lower levels of molecular weight, which is consistent with previous studies (Jamieson et al., 2014). FTIR analysis showed similar weak absorbance for both species and decay stage, with most noteworthy peaks present in all samples. Absorption around 1000 cm-1 is characteristic for C-O-C vibrations and represents aromaticity similar as measured with

UV/VIS (Norwood et al., 2013). More informative characterization of DC and insight in composition of DC could be achieved by applying 13_{C-NMR (Weishaar et al., 2001).}

Beech showed significant correlation with wood density for each characterization, whereas this was not the case for spruce. This could partly be attributed to the smaller range of wood density values for spruce, and hence more similarity over the decay continuum compared to beech. This distinction between wood of angiosperms and that of gymnosperms was previously found by Weedon et al. (Weedon et al., 2009), who also did not find significant relationships with wood traits for gymnosperm wood, like DOC concentration and aromaticity. Although decay stage showed a strong relationship with increased flammability, this did not show up as strongly when looking at the characterization of the DC. Only for beech a significant relationship with decay stage emerged. The minor differences in compositional characteristics can partly be attributed to the natural fire conditions under normal atmospheric conditions during the production of char. The maximum reached temperature was substrate dependent in this study, whereas many studies produce char at fixed temperatures in an oven. Variable temperatures during the laboratory fire reflected natural wildfire conditions, and resulted in rather minor temperature differences, the main driver of char properties (Gundale & DeLuca, 2006).

Microbial metabolism as a result of dissolved char

Microbial activity was measured both as respiration as well as microbial biomass, the spruce samples showed higher activity levels compared to the beech samples for both measurements. Respiration rate showed an exponential decay, with cumulatively more respiration for spruce. Wood density affected respiration rates for both species, but in opposing ways. Further stage of decay resulted in more cumulative respiration for spruce and less cumulative respiration for beech. The latter can partly be explained by the declining DOC levels with decay observed for the beech samples. Microbial biomass fluctuated during the first days whereafter it stabilized until the 18th _{day when it started to} increase consistently. Cumulative biomass levels differed between species, with the highest levels for spruce. No significant effect of decay stage for either species was observed. Biomass levels showed an increase during the 2nd _{half of the incubation experiment, with the most pronounced increase for the} freshly dead to partly decomposed branches compared to the very decomposed branches. This suggests an adaptation of the microbial communities towards using the more complex C molecules as energy source. A longer incubation experiment could give insight whether this trend persists on the long term. Microbial metabolism was assessed as the relative shifts in CUE-index. Respiration: biomass ratio yields an index of CUE, with a decreasing index indicating a relative increase in CUE. A drawback of this approach is that small fluctuations at very low biomass levels disproportionally influence the CUE-index. This was the case during the first 4 days of the incubation, which was therefore omitted from the analyses. An overall increase in CUE over time was observed for both species, with no significant effect for species or decay stage. This indicates that an adaptation towards the usage of char took place,

16 however the effect of either species or decay stage was not captured in this study. The relative activity levels for the spruce levels was higher, however this effect leveled out when looking at the CUE-index. The increased CUE over time indicates an adaptation of the microbial communities towards using char C as energy source. It is not likely that changes in microbial composition induced these changes, since previous studies found that the incorporation of biochar influenced soil microbial composition only after incubation periods extending 1 year (Jiang et al., 2016).

Synthesis

This study showed that microbial communities increase their CUE in a 4-week period with the addition of dissolved char irrespective of tree species. DC originates from pyrogenic material remaining on the forest floor after a forest fire. If microbial communities increase their efficiency, the C from the pyrogenic material can partly be stored in the soil in the form of microbial C according to the theory of Cotrufo et al. (Cotrufo et al., 2013). Furthermore, the increase in microbial derived compounds further promotes aggregation and the formation of organic-mineral bindings, the main mechanisms responsible for the formation of SOM. This implies that the rapid C release associated with forest fires can partly be mitigated by subsequent enhanced microbial C storage in forest soils.

Studies researching the effects of char addition often use mechanically produced char, this is appropriate when the effects of char on agricultural yields are investigated, however, to assess the effects of forest fires an approach more closely related to the natural conditions would be preferred. This study took a novel approach in the procedure of both obtaining char as well as performing the incubation experiment. The laboratory fire resulted in substrate dependent peak temperatures and combustion under the presence of oxygen, compared to pyrolysis produced char with preset temperatures and absence of oxygen (Hilscher et al., 2009; Jiang et al., 2016). The creation of char during wildfires is an analogy to anthropogenic biochar interventions but these cannot be treated identically (Sohi et al., 2010). Comparison with (bio)char studies indicated differences in the composition of the resulting char which can be attributed to the different production techniques. Natural forest fires are associated with far more confounding factors which influence the whole process, this further stresses the need of field situation focused research in order to accurately assess the effects in regard to C fluxes. The results from this study indicate genuine effects of the addition of DC to microbial metabolism, however extensive research is needed to further investigate the underlying mechanisms. The acquired results are exploitative since methodology and analyses strategy were developed during the experimental phase. This study delivered a promising framework for the set-up of future research on the effect of DC to microbial metabolism without the interference of a soil system. There are several recommendations for future research concerning the experimental set-up. Increasing the number of samples would enhance the robustness of the statistical analysis. To further enlarge the observed trends distinct classes could be created with small spreading for statistical strength. Activity fluctuations during the first 24-hours were not captured in detail, increased measuring resolution could give insight in this initial activity. It is also recommended to extent the incubation period since microbial biomass levels started to increase significantly after 3 weeks whereas respiration rate remained constant, it would be interesting to see whether this trend persists. Especially, because one study found that the addition of solid char to soils have found to stimulate microbial activity significantly only after an incubation period of months rather than weeks (Ameloot et al., 2013; Jiang et al., 2016). The underlying mechanisms could also be magnified by increasing the C concentration of both DC as well as the inoculum. In this study an initial C concentration of 20 mg/L was used which declined over time due to the microbial activity. It is expected that the accuracy of the measurements of microbial biomass with the direct chloroform extraction procedure would increase when the C concentration is higher with larger differences between fumigated and non-fumigated samples. Finally, this study made use of an adapted fumigation extraction protocol with direct liquid extraction without K2SO4 as extractant. It is known that char has 1 to 3 times the sorbent capacity of organic compounds compared to SOM, which is also less reversible (Durenkamp et al., 2010). Extractant concentrations were found to influence extraction efficiency significantly for different soils and biochar types with standard fumigation extraction, indicating the complex nature of measuring microbial biomass in relation to highly sorptive char (Durenkamp et al., 2010). Therefore, the accuracy of the direct chloroform extraction method in liquid without this extractant should be addressed in future studies.

17

Conclusion

This study aimed to clarify which mechanisms underlie the interaction between wood-derived dissolved char and the functioning of microbial communities in forest soils in respect to the fate of C. A laboratory fire experiment showed an increase in flammability with increasing decay stage for both the coniferous and the deciduous species. The composition of the dissolved char did not differ strongly between the species or stages of decay, for beech a slight increase in complexity in char composition with decay stage was observed. Incubating the DC with a forest floor representative inoculum showed an increase in CUE over time for both species, indicating an adaptation of their metabolism. This suggests that the process of C sequestration upon wildfires could be enhanced by the observed increase of CUE of microbial communities after the addition of DC, quantification of this process is not yet possible. Since the anthropogenic produced char is an analogy to char remaining after a wildfire, this study presents a promising approach on investigating the effects of DC on microbial metabolism in forest soils under real field conditions. The results from this study indicated a genuine effect of the addition of DC on microbial metabolism, however extensive research is needed to further elucidate the underlying mechanisms.

18

Acknowledgements

I would first like to thank my supervisor Albert Tietema for the support, assistance and valuable discussions during my whole thesis project. He guided me through designing and conducting an experiment from the beginning to the end and helped me to keep focus on the message.

I would also like to thank Hans Cornelissen and Richard Logtestijn for letting me use the Fire Laboratory of the Research in Ecology department of the Vrije Universiteit Amsterdam. Their knowledge on conducting fire experiments was essential for the successful performance of my experiment. Finally, I would like to thank the technicians from the IBED laboratory, and especially Rutger van Hall and Jorien Schoorl for their extensive assistance with both the set-up of the procedures as well as their practical help.

19

References

Ameloot, N., Graber, E. R., Verheijen, F. G. A., & De Neve, S. (2013). Interactions between biochar stability and soil organisms: review and research needs. European Journal of Soil Science, 64(4), 379–390. Bantle, A., Borken, W., Ellerbrock, R. H., Schulze, E. D., Weisser, W. W., & Matzner, E. (2014). Quantity and quality of dissolved organic carbon released from coarse woody debris of different tree species in the early phase of decomposition. Forest Ecology and Management, 329, 287–294. Berhe, A. A., & Kleber, M. (2013). Erosion, deposition, and the persistence of soil organic matter: mechanistic considerations and problems with terminology. Earth Surface Processes and Landforms, 38(8), 908–912. Błońska, E., Lasota, J., & Piaszczyk, W. (2018). Dissolved carbon and nitrogen release from deadwood of different tree species in various stages of decomposition. Soil Science and Plant Nutrition, 65(1), 100–107. Chave, J. (2005). Measuring wood density for tropical forest trees. A Field Manual for the CTFS Sites, 1–7. Cheng, C.-H., Lehmann, J., Thies, J. E., & Burton, S. D. (2008). Stability of black carbon in soils across a climatic gradient. Journal of Geophysical Research: Biogeosciences, 113(G2), n/a-n/a. Cornwell, W. K., Cornelissen, J. H. C., Allison, S. D., Bauhus, J., Eggleton, P., Preston, C. M., … Zanne, A. E. (2009). Plant traits and wood fates across the globe: rotted, burned, or consumed? Global Change Biology, 15(10), 2431–2449. Cotrufo, M. F., Wallenstein, M. D., Boot, C. M., Denef, K., & Paul, E. (2013). The Microbial Efficiency-Matrix Stabilization (MEMS) framework integrates plant litter decomposition with soil organic matter stabilization: do labile plant inputs form stable soil organic matter? Global Change Biology, 19(4), 988–995. de Nijs, E. A., & Cammeraat, E. L. H. (2020). The stability and fate of Soil Organic Carbon during the transport phase of soil erosion. Earth-Science Reviews, 201. de Nijs, E. A., Hicks, L. C., Leizeaga, A., Tietema, A., & Rousk, J. (2018). Soil microbial moisture dependences and responses to drying-rewetting: The legacy of 18 years drought. Global Change Biology, 25(3), 1005–1015.

Diffenbaugh, N. S., Swain, D. L., & Touma, D. (2015). Anthropogenic warming has increased drought risk in California. Proceedings of the National Academy of Sciences of the United States of America, 112(13), 3931–3936. Dungait, J. A. J., Hopkins, D. W., Gregory, A. S., & Whitmore, A. P. (2012). Soil organic matter turnover is governed by accessibility not recalcitrance. Global Change Biology, 18, 1781–1796. Durenkamp, M., Luo, Y., & Brookes, P. C. (2010). Impact of black carbon addition to soil on the determination of soil microbial biomass by fumigation extraction. Soil Biology and Biochemistry, 42(11), 2026–2029. Fellman, J. B., D’Amore, D. V., Hood, E., & Boone, R. D. (2008). Fluorescence characteristics and biodegradability of dissolved organic matter in forest and wetland soils from coastal temperate watersheds in southeast Alaska. Biogeochemistry, 88(2), 169–184. Freschet, G. T., Weedon, J. T., Aerts, R., van Hal, J. R., & Cornelissen, J. H. C. (2012). Interspecific differences in wood decay rates: insights from a new short-term method to study long-term wood decomposition. Journal of Ecology, 100(1), 161–170. Gregorich, E. G., Wen, G., Voroney, R. P., & Kachanoski, R. G. (1990). Calibration of a rapid direct chloroform extraction method for measuring soil microbial biomass C. Soil Biology and Biochemistry, 22(7), 1009–1011. Gundale, M. J., & DeLuca, T. H. (2006). Temperature and source material influence ecological attributes of ponderosa pine and Douglas-fir charcoal. Forest Ecology and Management, 231(1–3), 86–93.

He, Z., Mao, J., Honeycutt, C. W., Ohno, T., Hunt, J. F., & Cade-Menun, B. J. (2009). Characterization of plant-derived water extractable organic matter by multiple spectroscopic techniques. Biology and Fertility of Soils, 45(6), 609–616. Helms, J. R., Stubbins, A., Ritchie, J. D., Minor, E. C., Kieber, D. J., & Mopper, K. (2008). Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter. Limnology and Oceanographyfile:///Users/Evydenijs/Downloads/Helms_et_al-2008-Limnology_and_Oceanography.Pdf, 53(3), 955–969. Hilscher, A., Heister, K., Siewert, C., & Knicker, H. (2009). Mineralisation and structural changes during the initial phase of microbial degradation of pyrogenic plant residues in soil. Organic Geochemistry, 40(3), 332–342. Hirsch, P. R., Gilliam, L. M., Sohi, S. P., Williams, J. K., Clark, I. M., Murray, Phil J, & Murray, P J. (2009). Starving the soil of plant inputs for 50 years reduces abundance but not diversity of soil bacterial communities. Soil Biology and Biochemistry, 41, 2021–2024. Houghton, R. A. (2007). Balancing the Global Carbon Budget. Annual Review of Earth and Planetary Sciences, 35(1), 313–347. IPCC. (2013). Climate Change 2013 - The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, Thomas F. Qin, Dahe Plattner, Gian-Kasper Tignor, Melinda M.B. Allen, Simon K. Boschung, Judith Nauels, Alexander Xia, Yu Bex, Vincen. Jamieson, T., Sager, E., & Guéguen, C. (2014). Characterization of biochar-derived dissolved organic matter using UV–visible absorption and excitation–emission fluorescence spectroscopies. Chemosphere, 103, 197–204. Jiang, X., Denef, K., Stewart, C. E., & Cotrufo, M. F. (2016). Controls and dynamics of biochar decomposition and soil microbial abundance, composition, and carbon use efficiency during long-term biochar-amended soil incubations. Biology and Fertility of Soils, 52(1), 1–14. Kilpeläinen, A., Torssonen, P., Strandman, H., Kellomäki, S., Asikainen, A., & Peltola, H. (2016). Net climate impacts of forest biomass production and utilization in managed boreal forests. GCB Bioenergy, 8(2), 307–316. Knicker, H. (2007). How does fire affect the nature and stability of soil organic nitrogen and carbon? A review. Biogeochemistry, 85(1), 91–118. Lian, F., & Xing, B. (2017). Black Carbon (Biochar) In Water/Soil Environments: Molecular Structure, Sorption, Stability, and Potential Risk. Environmental Science & Technology, 51(23), 13517–13532. Lowden, L., & Hull, T. (2013). Flammability behaviour of wood and a review of the methods for its reduction. Fire Science Reviews, 2(4).

20

Magnússon, R., Tietema, A., Cornelissen, J. H. C., Hefting, M. M., & Kalbitz, K. (2016). Tamm Review: Sequestration of carbon from coarse woody debris in forest soils. Forest Ecology and Management, 377, 1–15.

Masiello, C. A. (2004). New directions in black carbon organic geochemistry. Marine Chemistry, 92(1–4), 201–213.

Norwood, M. J., Louchouarn, P., Kuo, L. J., & Harvey, O. R. (2013). Characterization and biodegradation of water-soluble biomarkers and organic carbon extracted from low temperature chars. Organic Geochemistry, 56, 111–119. Orth, R., Zscheischler, J., & Seneviratne, S. I. (2016). Record dry summer in 2015 challenges precipitation projections in Central Europe. Scientific Reports, 6. Regelink, I. C., Stoof, C. R., Rousseva, S., Weng, L., Lair, G. J., Kram, P., … Comans, R. N. J. (2015). Linkages between aggregate formation, porosity and soil chemical properties. Geoderma, 247–248, 24–37. Saiz, G., Wynn, J. G., Wurster, C. M., Goodrick, I., Nelson, P. N., & Bird, M. I. (2015). Pyrogenic carbon from tropical savanna burning: production and stable isotope composition. Biogeosciences, 12(6), 1849–1863. Schmidt, M. W. I., Torn, M. S., Abiven, S., Dittmar, T., Guggenberger, G., Janssens, I. A., … Trumbore, S. E. (2011). Persistence of soil organic matter as an ecosystem property. Nature, 478(7367), 49–56. Setia, R., Verma, S. L., & Marschner, P. (2012). Measuring microbial biomass carbon by direct extraction – Comparison with chloroform fumigation-extraction. European Journal of Soil Biology, 53, 103–106. Sohi, S. P., Krull, E., Lopez-Capel, E., & Bol, R. (2010). A review of biochar and its use and function in soil. In Advances in Agronomy (1st ed., Vol. 105, pp. 47–82). Elsevier Inc. Stockmann, U., Adams, M. A., Crawford, J. W., Field, D. J., Henakaarchchi, N., Jenkins, M., … Zimmermann, M. (2013). The knowns, known unknowns and unknowns of sequestration of soil organic carbon. Agriculture, Ecosystems and Environment, 164, 80–99. Weedon, J. T., Cornwell, W. K., Cornelissen, J. H. C., Zanne, A. E., Wirth, C., & Coomes, D. A. (2009). Global meta-analysis of wood decomposition rates: a role for trait variation among tree species? Ecology Letters, 12(1), 45–56. Weishaar, J., Aiken, G., Fram, M. S., Bergamaschi, B. A., & Fujii, R. (2001). Evaluation of specific UV absorbance as an indicator of the chemical content of dissolved organic carbon. Abstracts of Papers American Chemical Society, 222(1–2), ENVR199– ENVR199. Whitman, T., Hanley, K., Enders, A., & Lehmann, J. (2013). Predicting pyrogenic organic matter mineralization from its initial properties and implications for carbon management. Organic Geochemistry, 64, 76–83. Williamson, G. B., & Wiemann, M. C. (2010). Measuring wood specific gravity…Correctly. American Journal of Botany, 97(3), 519–524. Zhao, W., Blauw, L., van Logtestijn, R., Cornwell, W., Cornelissen, J., Zhao, W., … Cornelissen, J. H. C. (2014). Interactions between Fine Wood Decomposition and Flammability. Forests, 5(4), 827–846. Zhao, W., van Logtestijn, R. S. P., van der Werf, G. R., van Hal, J. R., & Cornelissen, J. H. C. (2018). Disentangling effects of key coarse woody debris fuel properties on its combustion, consumption and carbon gas emissions during experimental laboratory fire. Forest Ecology and Management, 427, 275–288.

21

Appendix

a. FTIR peak identification

The columns show the functional groups identified in the spectra. The rows show the measured samples by ID in descending wood density order. S stands for a spruce sample, B stands for a beech sample. Table 1 FTIR peak identification of each sample, categorized for amount of absorption in 3 classes; weak (30%), medium (30 – 60%), strong (>60%).

22

b. Incubation graphs: respiration, microbial biomass & CUE index ratio

Figure 17 Respiration rate per spruce sample during the incubation experiment. Figure 18 Respiration rate per beech sample during the incubation experiment.

23 Figure 19 Microbial biomass per spruce sample during the incubation experiment.

Figure 20 Microbial biomass per beech sample during the incubation experiment.

24

Figure 21 CUE-index per spruce sample during the incubation experiment.

Figure 22 CUE-index per beech sample during the incubation experiment.