The impact of historical summer droughts on the resistance and resilience of microbial communities to moisture stress

(1)

A collaboration between the Institute of Biodiversity and Ecosystem Dynamics of the University of Amsterdam and

the section of Microbial Ecology of Lund University

The impact of historical summer droughts on the

resistance and resilience of microbial communities to

moisture stress

Evy de Nijs

University of Amsterdam

Bachelor Bèta-Gamma,

major Earth Sciences

Amsterdam, June 3

rd

₂₀₁₇

Internal supervisor: dr. A. Tietema

External supervisor: J. Rousk

(2)

1

Abstract

Changing precipitation patterns are predicted as a result of climate change. Soil microbial communities are sensitive to changes in soil moisture, which is pronounced when rewetting a dry soil resulting in a pulse of CO2 emission. The aim of this study was to contribute to the understanding of future ecosystem acclimation upon changing precipitation stress. This was done by investigating the effect of historical summer drought on the response behaviour of microbial communities in soils when subjected to cycles of drying and rewetting. Specifically, whether experimental drought conditions (1) induce resistance to lower moisture, (2) induce resilience of microbial process rates, and (3) improve microbial carbon use efficiency during perturbation events. Composite samples from the organic layer were collected from an experimental field site on an Atlantic heathland in Oldebroek, the Netherlands. Here an annual summer drought experiment has been active for more than 15 years. Initial and final moisture dependence was established for both control and drought treated soils. Throughout the experiment, the soils were subjected to three cycles of drying and rewetting. Respiration rates and bacterial growth were measured during 75 hours after rewetting. The results show that soils subjected to drought in the field shifted towards higher drought resilience and resistance. Laboratory drying and rewetting resulted in increased resilience, but not resistance. Both field drying and laboratory drying perturbation resulted in communities with higher carbon use efficiencies. This study showed how soil microbial communities from an organic horizon respond to drought and cycles of drying and rewetting, thereby it contributes to the understanding of how ecosystems might respond upon changing precipitation patterns.

Keywords:

drying-rewetting, bacterial growth, respiration, organic soil, microbial communities, historical drought, drought adaptation, climate change

(3)

2

1. Introduction

Climate on earth is changing, driven by increased atmospheric concentrations of greenhouse gasses (IPCC, 2013). According to the latest IPCC report it is extremely likely that anthropogenic forces, and thus the influence of humans, have been the most important causes for global warming since the 1950s, but also on other effects of climate change the human influence can be detected (IPCC, 2013). A major contributor to the carbon flow from the terrestrial system to the atmosphere is soil respiration (Schimel, 1995). Soil microbial communities are sensitive to changes in environmental factors (Fierer & Schimel., 2002). The IPCC (2013) prediction is that the changes in the global water cycle will result in bigger contrasts in precipitation between regions both spatially and in time. Consequently, more extensive periods of drought followed by heavy precipitation events can be expected which will influence soil respiration and thus CO2 fluxes (Reichstein et al., 2013). Understanding the effects of increased intensity and periodicity of precipitation is critical to achieve a predictive understanding of how climate change might affect microbial communities, and thus regulate the carbon cycle of the terrestrial biosphere.

Microbes are the most abundant organisms in soils and highly dependent on the presence of water since it is their transporting medium, at the same time excessive amounts of water decreases microbial activity by the limitation of oxygen supply (Hueso et al., 2012; Pulleman & Tietema, 1998). The relationship between microbial communities and moisture content in soil is a complex balance. Several researchers have found that the fluctuations of soil moisture affect microbial communities on their growth rates, size and composition (Kim et al., 2012). Over 60 years ago, Birch (1958) was one of the first scientists to highlight the phenomenon that rewetting dry soil results in a pulse of respiration which are found to increase up to 4400% compared to the control soils (Kim et al., 2012). There are several proposed explanations for this sudden increase in respiration. One of these is that it can be ascribed to the metabolization of dead microbial substrate or non-living soil organic substrate released through the disturbance of drying and rewetting (Pulleman & Tietema, 1998; Fraser et al., 2016). However, many other explanations exist and without consensus it is not possible to pin down the cause of the occurrence of a respiration peak after rewetting. Over the years, a combination of laboratory and field investigations has given some new insights into the complex regulation of soil moisture fluctuations on microbial communities and processes.

Meisner and colleagues (2013) showed that bacterial communities from different soils can react in two fundamentally different ways to a rewetting event, named type 1 and type 2. The first response type can be characterized by an immediate linear increase in bacterial growth after rewetting which coincides with peak respiration rates just after rewetting followed by a continual decrease over time. The type 2 response can be characterised by a time lag with nearly zero bacterial growth, lasting for up to 20 hours, and followed by an exponential increase to very high rates. This growth response coincides with dynamics in respiration where rates initially are high and stable, sometimes followed by a secondary increase in rates in sync with growth rates

increases (figure 1). The mechanisms responsible for the occurrence of these different response types are not yet known, but recent insights have revealed that specific treatments can trigger changes between the two patterns. Meisner and colleagues (2015) found that for the researched soil a laboratory drought period ≤ 2 weeks resulted in type 1 responses, while periods ≥ 4 weeks

Figure 1; different response types upon rewetting (Rath et al., 2016)

(5)

4

resulted type 2 responses. Meisner and colleagues (2017) also found that with partial drying events a shift from a type 2 to a type 1 response might occur. Partial drying means that the soils are not completely air dried but kept at a higher moisture level. This indicates that, separated from the initial response type, also the history of drought exposure influences the response type of the microbes to different rewetting events. Thus, the harshness of the drying determines the type of bacterial growth response upon a rewetting event. When ecosystems are subjected to drought their response should be shifting, depending on their drought related history. Harsh drought perturbation will likely result in an overall more drought tolerant response.

Drought exposure results in a shift of the microbial community in the soil, evidence has been found for both selection for specific community compositions and enhanced CUE in these communities (Evans & Wallenstein, 2012; Göransson et al., 2013). Growing evidence suggests that the influence of historical conditions plays a major role on microbial responses to the environment (Fierer et al., 2003; Evans & Wallenstein, 2012). It has been shown that a community shift towards more stress tolerant taxa might occur even after a single drying and rewetting event (Evans & Wallenstein, 2012). Fierer and colleagues (2003) showed that microbial communities with a history of precipitation stress turned out to be more resistant when once more exposed to this disturbance compared to controls. The study of Evans and Wallenstein (2012) found that although soil moisture was the driving factor of microbial responses, the history regarding soil moisture also affects microbial responses to drying and rewetting events. This raises the theory that microbial communities which have been exposed to severe drought before might experience the new drying and rewetting as less of a perturbation compared to microbial communities which are not familiar with severe drought. If a soil has been exposed to drought and a microbial community shift has occurred, respiration rates are expected to be lowered since the Carbon Use Efficiency (CUE) of these microbes is expected to be enhanced, which means that their respiration per growth ratio is expected to be higher (Göransson et al., 2013). If this is combined with the findings of Meisner et al., (2017) it would be expected that if a soil shows an initial type 2 response this would shift to a type 1 response during several drying and rewetting cycles due to community shifts. Through this shift, communities will experience every subsequent drought stress event as a less harsh perturbation.

It has been shown that shifts in the composition of both bacterial and fungal communities can be related to changes in soil properties and thus play important roles in soils (Lauber et al., 2008). Previous research has shown that bacteria and fungi respond in a similar way to drying and rewetting disturbance, although the fungal responses are slightly slower and less pronounced (Meisner et., 2013). Therefore, it was chosen to put the focus of this study on the bacterial growth and respiration rates.

The aim of this research was to investigate how experimental field droughts influence microbial communities and their responses to rewetting events, contributing to the understanding of possible processes and feedback loops in relation to climate change. To achieve this, a soil was collected from a long-term field experiment located in Oldebroek, the Netherlands, on an Atlantic Heathland. Since the Oldebroek site has a history of 19 years of experimental annual summer drought exposure the expectation was that microbial communities might have formed which are more tolerant to drought and cycles of drying and rewetting. The objective of this research project was to test recent insights obtained from laboratory experiments in full scale field experiments. Specifically, whether experimental drought conditions (1) induces tolerance to lower moisture, (2) induces faster recovery to rewetting of microbial process rates, and (3) improves microbial C budget during perturbation events. The study addressed the following research question;

Does a history of summer drought change the resistance and resilience of microbial

communities to moisture and drying and rewetting events?

(6)

5

This research question was answered in a two-tiered approach. First, the effect of the summer drought field experiment was characterised in laboratory assessments. Secondly, the susceptibility of microbial communities to further cycles of drying and rewetting was investigated. Therefore, several hypotheses were tested. In this context, resistance refers to how well the microbial communities can handle drought and resilience refers to the speed of recovery after drought and the type of responses upon rewetting shown by the microbial communities.

It was hypothesized that (1) field experimental exposure to long-term drought would change microbial communities to (1a) grow more resistant to drought and (1b) respond with higher resilience to rewetting after drought. It was also hypothesized that (2) exposure to drying and rewetting (D/RW) would shift communities in the same directions as achieved in the field experiment, and that the initial differences between the field treatments would be maintained. The third hypothesis was that (3) both field drought and microcosm D/RW cycles would select for communities with a higher CUE during the perturbation cycles. A series of comparisons was used to evaluate these hypotheses.

The outcome of the field experiment (1) was tested by establishing the moisture dependence (1a) and examining the responses to the first perturbation cycle of drying and rewetting (1b). The hypothesis was that the microbial communities of the ‘drought’ treated soils have shifted to be more drought tolerant than the ‘control’ soils and were thus more resistant to drought. For the moisture dependence, this would result in higher levels of activity, for both respiration rates and microbial growth, at lower moisture levels for the drought treated soils compared to lower levels of activity for the control treated soils (figure 2a). For the first drying and rewetting cycle, the expectation was that drought treated soils would experience the perturbation as less harsh than control soils. If the soil showed an initial type 2 response the two treatments would be tested by comparing the observed time lags for bacterial growth, which was hypothesized to be shorter for the drought treated soils (figure 2b). For respiration, lower respiration rates and thus a lower total respiration was hypothesized since the CUE of the drought treated microbes were expected to be enhanced compared to the control treated microbial communities (figure 2c).

Figure 2; schematic representations supporting the hypotheses, the arrow points towards drought treated soil responses A) moisture dependence B) microbial growth upon rewetting C) respiration upon rewetting

(7)

6

The second hypothesis (2) was tested by exposing the communities to several perturbations of drying and rewetting to test the resilience. After all three cycles, moisture dependence was re-established to check for changes in resilience. The expectation was that the microbial communities would adapt to be more drought tolerant over the course of the different cycles. If the soil showed an initial type 2 response it was expected that over the course of one to several drying and rewetting cycles this would change into a type 1 response due to community shifts. Regarding the final moisture dependence, the hypothesis was that both soils would become more adapted to harsh perturbation and thus maintain their activity rates longer at lower temperatures, while the initial differences would be maintained.

To evaluate the hypothesis (3) regarding selection for carbon efficient communities upon drought stress due to community shifts, the cumulative respiration to bacterial growth ratios were compared. The expectation was that the ratio of cumulative respiration rates to the bacterial growth of the drought treated soils in the first cycle would be lower compared to the control treated soils. And that this ratio would, for both treatments, decrease over the different cycles due to selection for communities with a higher CUE.

2. Materials and methods

2.1 Field experiment

The soil samples were taken from an experimental field treatment site in Oldebroek, the Netherlands (52° 24' N, 005° 55' E). An experimental summer drought treatment has been active on three plots since 1998, which resulted in sequence of summer droughts. During 2-5 months in the summer the drought treatment was activated. The drought treatment is active by a retractable transparent and waterproof cover which masks the plots during rain events which last longer than 30 minutes, preventing rain to enter the soil (Beier et al., 2004). Also, three control plots with the same framework but without any active treatment were located at this field site. The effect of the drought treatment at the Oldebroek experimental site was an approximate precipitation reduction of 31% during the growing season (Rousk et al., 2013). For this research, samples from both plot treatments were used, thus one set received the experimental prolonged summer drought where the other set did not receive a treatment.

2.2 Soil samples

Soil was collected at the end of April 2017 from the experimental site in Oldebroek. It is a well-drained grassland soil, classified as a Haplic arenosol (FAO). A total of six samples was taken from the experimental field treatments ‘drought’ and ‘control’, each with three replicates. Composite samples from the pure organic matter horizon were sampled from both the historical drought treatment plots and the control plots (n=3). Approximately 6 to 8 cores per plot were taken and the organic matter horizon had a thickness varying between 2 and 8 cm. This resulted in a total of 6 samples with an approximate 300-gr dry weight equivalent. Immediately after taking the samples, big components such as branches, stones and cones were removed from the samples and in the laboratory the samples were homogenized by hand and wet-sieved (<6 mm). The soil samples were stored airtight and dark at temperatures below 8 °C until the start of the experiments (ca. 5 days).

Figure 3; schematic representation of a shift in response type upon rewetting a dried soil, supporting hypothesis 2

(8)

7

2.3 Soil characteristics

The following soil characteristics were assessed after sampling: water content (gravimetric, 24 h at 105 °C), soil organic matter (loss on ignition, 12 h at 550 °C), soil pH & conductivity (1:5, soil: distilled water extraction). The Water Holding Capacity (WHC) of each soil was determined by letting the soils soak up water for at least 12 hours and let them drain for at least 6 hours. Basal respiration and bacterial growth were determined using the techniques described below (2.6). Basal fungal growth was determined using 14_{C-acetate incorporation into ergosterol as described} by Bååth (2001). Except for pH, conductivity and WHC the measurements where done using two technical replicates, the average of these were used for further analyses.

2.4 Moisture dependence

Moisture dependence of each of the samples was determined immediately after sampling and after the three drying and rewetting cycles. Subsamples of ca. 50 grams soil at 50% WHC were put in microcosms and placed under a ventilator at an average temperature of 24°C to dry. Every 1 to 2 hours, samples of 0.5 grams were taken for respiration and microbial growth and a 1 gram sample for water content. This was continued until ca. 24 hours after starting the drying process when the moisture content stabilized. The initial moisture dependence was established on the soils after sampling. The moisture dependence after the different D/RW cycles was performed on both the cycled soils as well as 50% WHC continuously moist soils, which were kept at 4°C when not used for measurements to minimize the effect of change due to storage. The techniques used for analysing respiration and bacterial growth are explained in section 2.6.

2.5 Experimental D/RW cycles

The samples from both treatments were exposed to three cycles of drying and rewetting, further stated as D/RW cycles (figure 4). First, fresh soil of each of the samples was weighed into plastic containers with lids and adjusted to 50% water holding capacity (WHC) by gravimetrical weight. These were incubated for 2 days to stabilize. For the D/RW cycles, controls for each replicate which did not receive the cycling treatment were kept at 4°C during the experiment. The control of each replicate was analysed for respiration and bacterial growth at several time points during the cycles. This control data was used for replicate normalisation.

(9)

8

The microcosms, with the lids off, were placed under a ventilator at a stable 24 °C temperature to speed up the drying process. The establishment of the initial moisture dependence showed that after 24 hours under the ventilator the soils reached a stable moisture level. Therefore, a minimum of 24 hours under the ventilator was needed for complete air drying, in practice the drying was done over the weekend which resulted in a period of approximately 65 hours. The mean moisture content for the air-dried ‘control’ and ‘drought’ samples was ca. 2.15% WHC. For rewetting, the moisture levels of all soils were increased to their 50% WHC. The samples were rewetted using a pipette with demineralized water and mixed using a spatula to make sure the moisture was evenly distributed within each sample. After rewetting all samples were stored in a stable 16 °C room until the subsamples of the specific timepoint were measured.

Respiration rates and bacterial growth were measured until approximately 4 days after rewetting since previous research has shown that this period covers the increase in activity upon rewetting a dried soil. To assure a high time resolution, a morning and evening set was prepared for each soil and rewetted in either the morning or the evening. For respiration, 7 measurements were made, starting with an interval of 6 hours until 24 hours after rewetting followed by a measuring interval of 12-24 hours. The length of these intervals corresponded with the time the samples were incubated for. For bacterial growth, a total of 18 measurements were made with an interval of 3 hours for the first 36 hours after rewetting, after that an interval of 6 hours increasing to 12 hours in the last 2 days was maintained (appendix 8.1). The continuously moist controls (CMC) of both respiration and microbial growth were measured 3 times during each cycle, this data was used for replicate normalisation.

2.6 Measurements

2.6.1 Respiration

The accumulated CO2 was measured for a certain time block in the period after rewetting to measure the respiration. 0.5-gram subsamples were placed in glass vials, purged with pressured air subsequently sealed and stored for incubation in a dark and temperature stable room of 16°C. After a pre-determined incubation time, the CO2 concentration was analysed using a gas chromatograph equipped with a flame ionization detector and a methanizer. Respiration was expressed as µg CO2 per gram dry weight soil per hour.

2.6.2 Bacterial growth

Bacterial growth was measured using a leucine incorporation method according to Bååth et al. (2001). The incorporation of 3_{H-labelled leucine into extracted soil bacteria was used as a relative} measure of microbial growth. 0.5-gram subsamples of soil were mixed with 20 ml demineralized water and vortexed for 3 minutes followed by 10 minutes of centrifuging at 3000 rpm. This resulted in a bacterial suspension of which 1.5 ml subsamples were put in Eppendorf vials. 10 µl of 3_{H-Leucine solution (2µl}3_{H-Leu, 2 µl 1:9 Leu, 6 µl dH2O) was added to each Eppendorf vial. The} vials were incubated for 1 hour in a dark stable 16°C room before the microbial incorporation was ended by adding 100% TCA (trichloro-acetic acid) which is a lethal solution for microbes. Several steps of washing and preparing for scintillation counting were performed as described by Bååth and colleagues (2001). This resulted in an amount of incorporated leucine into bacteria per gram dry weight soil per hour, and used as a proxy for bacterial growth.

2.7 Statistical analysis

Kaleidagraph 4.5.3 (Synergy Software) and Matlab2016b were used for the visualisation and analyses of the data. When curves were fitted through the data, the R2_{-value was used as a} measurement of suitability.

(10)

9

The moisture dependence data as well as the bacterial growth of the first cycle was fitted using a sigmoidal curve fit (Rath et al., 2016). A logistic model with the following equation was applied:

𝑦 =_1+𝑒_{−𝑏(𝑥−𝑎)}𝑐 (1)

y describes the activity at a certain x-value (i.e. moisture level or hours after rewetting), a is the

50% inhibition value, b is a parameter which indicates the inhibition rate and c represents the process rate for the control (Rath et al., 2016). For the respiration data of the different cycles a standard exponential decrease fit was applied. The bacterial growth for cycle 2 & 3 was fitted using a linear fit until the point of substrate exhaustion followed by an exponential decrease curve. The bacterial growth data for the first cycle with the type 2 response was fitted with the Gompertz model (Meisner et al., 2015). This model uses the LOG transformed data to describe bacterial growth:

𝐺𝑡 = 𝐺𝑡0+ 𝐴 × 𝑒−𝑒 𝑏−𝑐𝑡

(2) 𝐿𝑎𝑔 𝑡𝑖𝑚𝑒 = 𝑏−1

𝑐 (3)

The Gt is the logarithm of the log transformed normalised bacterial growth data, A represents the

difference between the extreme values, b and c are fitted parameters. The time lag for bacterial growth is calculated according to equation (3).

The lag times of both treatments for each replicate were tested using a t-test, the IC50 values and basal treatment differences were tested with analyses of variance (ANOVA). For the fitted curves differences at a level of α=0.10 were held significant, otherwise α=0.05 was held as significance level.

All cumulative data was log transformed to minimize scaling effects in the variances. The cumulative values for the resilience outcome of the field experiment, thus the first cycle, were tested in a one-way ANOVA. This was chosen since the interest of the first hypothesis is only on the field effects and thus the first cycle, therefore it was not needed to take later cycles into account. The cumulative values of all cycles were tested with a n-way ANOVA, this was done for both 24 hours and the total period measured. Differences at level of α=0.05 were held significant. Due to a limited amount of repetitions per treatment per cycle it was not possible to test all interaction effects. A n-way ANOVA in combination with the post hoc Tukey HSD test was chosen since this takes repeated measurements on replicates into account.

(11)

10

3. Results

The graphs with raw data points for the initial moisture dependence can be found in appendix 8.2.1, for the cycles these can be found in appendix 8.2.2, finally the raw data graphs for the final moisture dependence can be found in appendix 8.2.3. Tables with the results of the ANOVA tests and Tukey HSD test can be found in appendix 8.4

3.1 Soil characteristics & basal activity

The basic soil characteristics for the 3 control treated soils (C1-C3) and for the 3 drought treated soils (D1-D3) were measured (table 1). No significant differences between the properties of the soils from the two field treatment types was detected.

Table 1; soil physiochemistry, the significance between the treatments was tested using a one-way ANOVA (α<0.05)

C1 C2 C3 D1 D2 D3 SIGNIFICANT? WATER CONTENT (%) 199 133 230 172 165 124 No (p=0.36) SOM (%) 76.5 55.1 85.1 71.5 77.3 75.5 No (p=0.79) PH 1:5 EXTRACTION 4.3 4.3 4.3 4.5 4.4 4.1 No (p=0.72) CONDUCTIVITY (MS M-1₎ 1:5 EXTRACTION 7.6 3.3 4.5 4.1 3.2 6.6 No (p=0.76) The average Water Holding Capacity (WHC) for the control soils was 486±51 SE % and for the drought soils 521±28 SE %. A one-way ANOVA gave a p-value of 0.65, thus the WHC of the soils of the two treatments was not significantly different. The basal bacterial growth for the control soils was on average 194±24 SE pmol Leu/g/hr, for the drought soils this was 155±12 SE pmol Leu/g/hr, these were not significantly different (p=0.21). Basal respiration rates differed between the different individual soils, however the mean of the grouped control and drought soils were similar and non-significant (resp. 16±1 SE & 15±1 SE µg CO2/g/hr, p=0.72). The relative fungal growth was 465±59 SE pmol Ac/h for the control soils and 352±69 SE pmol Ac/g/h for the drought soils. Thus, the basal fungal growth was higher in the control samples than in the drought samples, however this observed difference was not significant (p=0.18).

(12)

11

3.2 Field experiment effects

3.2.1 Initial moisture dependence

For moisture dependence, both normalised bacterial growth and respiration were plotted against the % of water holding capacity (WHC) per treatment. A logistic curve was fitted through all raw data points of the replicates per treatment.

The respiration was the same for soil moisture levels above ca. 20% of the WHC, thereafter the activity, for both respiration and bacterial growth, was maintained longer for the drought treated soils (red line) compared to the control treated soils (blue line, figure 5). The differences in activity at certain moisture contents became less distinct when soil moisture, as % of WHC, drops below 10%. For bacterial growth, the differences between the two treatments was continually present (figure 6). The bacterial growth rates were maintained longer in the drought treated soils when moisture content decreases, compared to the control soils.

Figure 5; initial moisture dependence, respiration Figure 6; initial moisture dependence, bacterial growth

The IC50 (50% inhibition concentration) values per treatment of each replicate were tested in a one-way ANOVA. Testing the IC50 values for respiration gave a p-value of 0.2602, thus these were not significantly different. The IC50 values for bacterial growth are significantly different between the control and drought treated soils (p<0.1). The 50% bacterial growth activity of the microbes in the drought treated soils showed a trend to occur at a lower moisture level compared to the control soils.

(13)

12

3.2.2 Bacterial growth & respiration

The Oldebroek soil showed a type 2 response upon rewetting. The respiration remained relatively high for a long period (figure 7) and bacterial growth started increasing exponential after a time lag (figure 8). The fitted curve for the drought treated soils (red) was shifted downwards for respiration and was shifted to the left for bacterial growth compared to the curve for the control soils (blue).

Figure 7; D/RW1, respiration Figure 8; D/RW1, bacterial growth

A tendency in decreased respiration for the drought treated soils was observed visually. The differences in cumulative values were tested in a one-way ANOVA. The respiration for the drought and control treated soils were cumulatively seen not significantly different for either the first 24 hours or the total period (p=0.595 & p=0.682). The average cumulative values of bacterial growth for the first 24 hours was 712±17 SE pmol Leu for the control soils and 1038±25 SE pmol Leu for the drought soils. This difference was tested significant (p<0.001), thus cumulative bacterial growth is significantly higher for the drought compared to the control soils for the first 24 hours. For the total measured period, cumulative bacterial growth was not significantly different between the treatments. For bacterial growth, the average time lag for the drought soils was 6.2 hours compared to 8.6 hours for the control soils. The individual replicate values of each treatment were tested in a t-test, which resulted in a trend with a p-value of 0.057 and thus bordering on significance. The respiration: bacterial growth ratio for the control soils was significantly (p<0.05) higher (1.58±0.09 SE) than the ratio for the drought treated soils (0.98±0.09 SE).

(14)

13

3.3 Cycles of drying and rewetting

3.3.1 Respiration

The respiration data was modelled with exponential decay curves for all cycles. The respiration graph of the first cycle has previously been discussed in section 3.2.2. No significant effects in cumulative respiration between the treatments were present, the effect of cycling was significant for both treatments (p=0). No interaction effects between treatment and cycles were tested significant. Respiration in the second cycle started slightly lower and showed a higher exponential decrease quickly after rewetting compared to the first cycle (figure 9B). A cumulative decrease, for the first 24 hours and the total period, between the first and second cycle was for both treatments significant (p<0.05). The third cycle started even lower and subsequently followed a similar response type as the second cycle (figure 9C). No significant differences in a cumulative decrease of respiration between the second and third cycle for either of the field treatments was present.

(15)

14

3.3.2 Bacterial growth

Bacterial growth was modelled with a logistic curve for the first cycle, for the following two cycles a combination of a linear curve for the first 48 hours for cycle 2 and 30 hours for cycle 3 was followed by an exponential curve for the second part of the graph. The isolated effects of the first cycle have previously been discussed in 3.2.2. Isolating the treatments in the first cycle in a one-way ANOVA resulted in significant (p<0.001) increased bacterial growth for drought for the first 24 hours (figure 10A). However, including both treatment and cycling as factors in a n-way ANOVA did not result in a significant difference. As previously discussed, a trend in shorter time lag for the drought treated soils was present (p<0.06). After the second dry down the response upon rewetting showed a type 1 response, with a linear increase immediately upon rewetting and a later decrease (figure 10B), this response type is maintained in the third cycle (figure 10C). The effect of cycles had a significant influence on both time frames (p<0.05). The effect of treatment was not found significant, however it bordered on significance for the first 24 hours (p=0.057). No significant interaction effects were observed. Cumulative bacterial growth for the total period decreased between the first and second for the drought treated soils (p=0.040). No significant differences between the treatments within cycles was observed. The bacterial growth in cycle 1 stabilized after ca. 30 hours, in cycle 2 after ca. 48 hours and in cycle 3 after ca. 30 hours.

(16)

15

3.3.3 Cumulative values & ratios

The cumulative respiration rates and microbial growth for the first 24 hours and 75-hour measured period were calculated and standardized for comparison, using the within replicate moist control values (control = 1). The error bars represent standard errors. The effect of cycling was for both timeframes and measurements significant. For the first 24 hours, the effect of treatment was significant (p=0.012), the interaction between the two factors was not found significant (p=0.063).

For the first 24 hours and the total period, cumulative respiration decreased for both control and drought significantly (p<0.05) between the first and second cycle (figure 11 & 12). Cumulative bacterial growth was neither between the cycles nor the treatments significantly different for the first 24 hours. Cumulative bacterial growth decreased significantly during the total experiment for the drought treated soils (p<0.05). The control soils were not significantly different between the first and second cycle (p=0.062), however a significant decrease between the first and third cycle was present (p=0.028). In other words, for the complete period the total respiration decreased after the first cycle for both treatments, after which these were maintained. In the same period, bacterial growth decreased significantly after the first cycle for drought and between the first and third cycle for the control treated soils (figure 12). For neither of the time frames, significant differences between the treatments within the cycles were observed.

Figure 11; cumulative values, 24 hours Figure 12; cumulative values, 75 hours

The ratios of respiration to bacterial growth for the different cycles and treatments were tested in a n-way ANOVA, thus considering all factors of the experiment. For the first 24 hours both factors (cycles & treatments) as well as the interaction were significant (p<0.02). For the two treatments, significant differences in the first cycle were present. The group cycle 1 control is significantly different from group cycle 1 drought (p<0.05). The ratio for the control soils was thus significantly higher (1.58±0.09 SE) than the ratio for the drought treated soils (0.98±0.09 SE). In the first cycle, the ratio respiration to bacterial growth was significantly higher compared to the other cycles, for both control and drought (p<0.002). Within and between cycle 2 & 3 no significant differences for cycle or treatment were present. For the total period, the factor cycle was bordering on significance (p=0.056), none of the ratios was significantly different from other groups.

(17)

16

3.3.4 Final moisture dependence

The final moisture dependence was established for the soils which received three cycles of drying and rewetting. Also, the moisture dependence of the continuously moist control soils was re-established to take away potential storage effects. For bacterial growth, the variation between the replicates was in such extent high that no valuable conclusion could be made (appendix 8.3). For respiration, no differences could be established between the different field treatments. However, the three cycles of D/RW clearly resulting in a shift in moisture dependence of both field experimental treatments. These shifts suggested that less drought resistant microbial communities had resulted from the repeated cycles of D/RW (figure 13). The IC50 levels of all soils after cycling showed a similar shift towards the right (p=0.0025). The same respiration occurred at higher soil moisture levels after cycling the soils compared to the moist control soils. Thus, the cycled soils became significantly less moisture tolerant due to the cycling in drying and rewetting.

(18)

17

4. Discussion

This experiment was conducted to investigate whether a history of summer drought would change the resistance and resilience of microbial communities to moisture and drying and rewetting perturbation. Subsequently, the evolution of drought tolerance was tested in a sequence of three cycles of drying and rewetting. The soil originated from a field experiment with historical summer drought and control plots, sampled prior to the annual start of the treatment for 2017. The soil characteristic analyses showed that the samples were on average high in organic carbon (>70%), and had similar values for pH and conductivity. The basal activity of the soils for respiration, bacterial & fungal growth seemed to be slightly lower for the drought treated soils, however these differences were not significant. Thus, the soil characteristics of all samples were similar.

4.1 Field experiment effects

It was hypothesized that repeated experimental summer drought exposure in the field would result in a change in microbial communities to (1a) grow more resistant to drought and (1b) respond with higher resilience to rewetting after drought. With the initial moisture dependence, the resistance of the communities was tested, with the first cycle of drying and rewetting the resilience of the communities was tested.

Moisture dependence showed that the microbial communities of the drought treated soils were more drought resistant regarding their bacterial growth compared to the control treated soils. For respiration, the stable level of activity is maintained longer at lower moisture levels compared to the control soils. Since the drought treated microbial communities are slightly better in managing drought, the effect of the long-term field experiment is that there is a trend in selection for more drought resistant microbial communities. The moisture dependence showed that respiration is maintained relatively stable until moisture levels of ca. 20% of WHC, whereas bacterial growth decreases continually with decreasing moisture levels. The process of respiration is thus less sensitive to moisture reduction than bacterial growth and the respiration per bacterial growth ratio increases when soil moisture levels decreases. This implicates that when ecosystems are exposed to partial drying the CUE of the microbial communities decreases and thus relatively more carbon is released during partial dry periods. However, this hypothesis is not supported by a study where it was observed that rewetting had no effect on respiration or growth when a threshold of ca. 30% WHC was crossed (Iovieno & Bååth, 2008; Meisner et al., 2017). Since the CUE for the moisture dependence is not explicitly studied here, no clarity on actual differences can be given. Another point to make here is that the respiration measurements for moisture dependence do not cover the activity at low moisture levels, which resulted in a jump in the measurements from high levels of activity to very low levels of activity.

The results of the first cycle of drying and rewetting were used to test differences in resilience between the two treatments. The soil of both treatments showed a type 2 response upon rewetting since the bacterial growth started increasing exponential after a time lag of respectively 6.2±1.0 SE and 8.7±0.8 SE hours for the drought and control treated soils. Meisner and colleagues (2015) found that a type 2 response occurred if a soil was dried for ≥4 weeks, since our soils showed a type 2 response after only 2.5 days of drying we can conclude that this soil has an initial type 2 response. The processes resulting in different response behaviours for soils for peculiar drying times remains to be resolved (Meisner et al., 2017). The shorter time lag for the drought treated soils does not stroke with the findings of Göransson et al., (2013), who looked at soils from an alder- and beech forest and found a type 2 response, but with a shorter time lag for the control treated soils compared to the soils which received a drought treatment for two summers. Cumulative bacterial growth in the first 24 hours was significantly higher for the drought treated communities compared to the control treated communities. Thus, the drought treated soils showed a trend towards decreased time lag before bacterial growth and had a higher cumulative

(19)

18

growth in the first 24 hours compared to the control groups. It was also hypothesized that (3) the field drought treatment would select for more carbon efficient communities during perturbation cycles. The ratio respiration to bacterial growth for the first 24 hours in the first cycle is significantly higher for the controls than for the drought treated soils (1.58 vs 0.98). A lower ratio means that the communities are more efficient using the available carbon resources, which in this case was found for the drought treated microbial communities. This supports the hypothesis that the summer drought field treatment selected for communities with a higher CUE. The study of Göransson and colleagues (2013) might have found contradicting results for time lags, but also found that exposure to drought resulted in microbial communities with higher CUE. This increase in microbial C use efficiency under drier conditions has also been observed in other studies where they increased the periodicity of precipitation (Zeglin et al., 2013), compared dry and wet months (Zeglin et al., 2013) and a study where they looked at microbial growth efficiencies along a moisture gradient (Herron et al., 2009). However, the opposite has also been observed in a study by Tiemann and Billings (2011) who found declining CUE with increased soil moisture stress in their in-situ study of a grassland system.

The microbial communities which received summer drought for the last 19 years are slightly more adapted to drought. A selection on both resistant and resilient microbial communities has occurred during the time the experimental field treatment has been active. The ratio respiration to bacterial growth also showed selection for a higher CUE during the drought treatment. These results are important for future predictions, since it suggests that microbial communities do adapt to changing climatic circumstances also when the changed climatic conditions are only present during summer.

4.2 Cycles of drying and rewetting

For the different cycles of drying and rewetting, it was hypothesized (2) that exposure to D/RW events would shift microbial communities to be more drought resistant and resilient, and that initial differences between the treatments would be maintained. Both respiration rates and bacterial growth where measured upon rewetting for the three perturbation cycles the soils were subjected to.

The bacterial growth in the first cycle showed a clear type 2 response, with a time lag after rewetting followed by an exponential increase (Meisner et al., 2013). The drought treated soils showed a trend towards increased drought resilience by having a shorter time lag compared to the control treated soils. After the first cycle, the response of both soils shifted from the type 2 response to a type 1 response. This was in line with the second hypothesis, since the soils were subjected to harsh drying which was expected to trigger a community shift. This relatively quick shift is in line with a high turn-over rate for soil microbial biomass, this results in communities which can change drastically to new conditions in a short timeframe (Lipson et al., 2002). The physical harshness of the drying and rewetting perturbation was the same over the different cycles however the community shift resulted in more tolerant microbes which thus experience the perturbations as less harsh. This is supported by work from Lipson and colleagues (2002), they found that microbial communities from an Alpine meadow underwent a shift in both structure and function, and thus community shifts occurred, between winter and summer, due to the sudden changes in temperatures and substrate availability. Along these lines microbial population dynamics depending on seasonal variability can partly be described by changes in community shifts.

The differences for bacterial growth between the treatments decreased over the cycles until being absent after the third cycle. The bacterial growth in the second and third cycle showed a type 1 response, with linear growth upon rewetting switching to a stable or decreasing growth after a certain period. The distinction between the two phases was based on a fitted logistic curve, which was not representable since two different processes were active but does give a clear division

(20)

19

between the phases (Zwietering et al., 1990). The switch occurred when the point of substrate exhaustion had been reached. Because consumption rates and thus availability of carbon matches growth, bacterial growth can no longer be sustained if carbon deficiency occurs. The growth rates for the first cycle stabilizes after ca. 30 hours, and the non-fitted data showed a decrease in growth after ca. 42 hours, supporting the theory of depletion of available carbon. The bacterial growth in the second cycle stabilizes after ca. 48 hours. This increase in depletion time can be explained by the process of shifting from a type 2 to a type 1 response where the total growth in the first cycle is higher than for the second cycle and thus a higher carbon demand after the time lag which leads to earlier depletion. The point of research exhaustion in the third cycle was reached after ca. 30 hours. Since the response type for the last two cycles was similar, this effect can be explained by the fact that the total carbon in the soils decreased over the different cycles. The physical disturbance associated with rewetting results in the release of accessible substrate in every cycle (Miller et al., 2005; Xiang et al., 2008). This decrease in carbon availability over the course of the different cycles is also visible in the respiration curves. Every cycle started with lower respiration rates for both treatments and where carbon wealth was maintained relatively long for the first cycle it was maintained briefer in the second cycle and yet briefer in the third. A declining rewetting effect has also been observed by Borken and Matzer (2009), where frequent drying and rewetting reduced the rewetting pulse due to limitations in freely available organic matter. The differences between the first and later cycles can partly be explained by the switch from a type 2 to a type 1 response. During the different cycles, respiration bended progressively more downwards for both treatments. These results match the results of the bacterial growth, which was expected since the respiration resembles energy sources and thus energy availability for growth. A similar decrease in respiration over the different cycles was observed in a study from Miller and colleagues (2005), however it is in contrast with a study from Xiang and colleagues (2008). The latter found increasing microbial respiration rates which were not in proportion with the total initial microbial biomass. In our study, respiration was 1.5 times higher after the third cycle compared to the moist control but started initially at levels 2.8 times the control respiration. This range of respiration increase due to cycling is in agreement with previous studies and the proposition that the history of rewetting does influence the magnitude of the respiration pulse (Miller et al., 2005; Fierer and Schimel, 2003).

It was hypothesized (3) that both field treatment and D/RW perturbations would select for communities with a higher CUE. The first cycle supported the hypothesis that the field treatment did select for communities with a higher CUE. The respiration to bacterial growth ratios showed that for both control and drought the ratios in the first cycle were significantly higher than in the following cycles. No differences within and between the second and third cycle were observed. Thus, during the cycling experiment only significant selection for carbon efficient communities occurred between the first and second cycle. This coincided with the transitions from a type 2 to a type 1 response. Fierer and colleagues (2003) found that the bacterial compositions regarding taxonomic diversity and richness was relatively insensitive to the cycles of D/RW. They proposed that the observed shifts, induced by D/RW, were originating from changes in carbon mineralization rates, which is in line with increased CUE observed in this research (Fierer et al., 2003). The effect of cycles is for both treatments and time frames a significant factor. Moreover, for the respiration per bacterial growth, both field treatment and cycle had a significant influence in the first 24 hours. These differences gradually disappeared with additional cycles since the interaction term was non-significant. The initial differences in the first cycle, and thus the field treatment effects are not maintained but instead disappear with additional cycles, this partly contradicts hypothesis 3.

After the three D/RW cycles, the moisture dependence of the cycled soils and the continually moist controls was re-established to test the changes in resistance of the microbial communities. It was hypothesized that the microbial communities of both treatment soils would become more resistant to drought since they were exposed to 3 drying perturbations where selective pressures might have been present. However, the moisture dependence showed that the microbial

(21)

20

communities of both soils became significantly less drought resistance after enduring three D/RW cycles. Apparently, no selection for drought resistant microbial communities has taken place. This raises the theory that the microbial communities did not experience the drying perturbation as harsh and thus no selective pressure occurred. It follows that the rewetting perturbation might have been the selective pressure, and so selection for resilient communities occurred. The cycles showed that the microbial communities became more resilient to drought since the response type shifted from type 2 to type 1. The initial moisture dependence gave insight in the selection on resistance during the field experiment. A trend for selection of drought resistant microbial communities was observed for the drought treated soils. This means that in the field, with partial drying, both selection for resistant and resilient microbial communities has taken place while during the ad hoc laboratory cycling experiment only selection for resilience within microbial communities took place.

4.3 Summer drought & cycles of perturbations

Harsh laboratory drying and rewetting cycles induced a shift from a type 2 response towards a type 1 response. A less pronounced, but comparable difference was found for the field treatments. The drought treated soils showed a less extreme type 2 response upon the first rewetting event, and thus were more resilient to drought. The same pattern was visible for CUE, both the field drought treatment and the cycles resulted in microbial communities which consumed carbon more efficiently. This means that one component that has been exerting a selective pressure on the microbial communities in the field is resilience against cycles of D/RW. Also, a slight increase in resistance to drought was found, however, the experiment showed a decreased drought tolerance after cycling, thus no selection on increased resistance against drought took place in the laboratory. In other words, the field experiment resulted in a slightly stronger resistance against drought whereas laboratory experiment did not. This must mean that the selective force shaping microbial communities in the field exerts a different type of pressure compared to the type of pressure applied in the laboratory. One explanation is that the drying conditions applied in the laboratory did not resemble the drying conditions in the field. In the laboratory, the soils were dried harshly to the level of air dry while it is presumable that soils in the field experienced partial drying most of the time. Meisner and colleagues (2017) have shown that the type of drying has a relative big influence on the response types and can even shift these. Subsequently, the hypothesis can be raised that in the field selection for different intensities of drought has taken place. Therefore, it would be interesting to repeat this experiment but then apply partial drying, thus drying to ca. 20% WHC. In addition to this, since no selection occurred during the perturbation cycles the idea was raised that the microbial communities experienced the rewetting event as a selective perturbation, and thus only selected for resilience. The resilience of the communities did increase since a shift from type 2 to type 1 occurred, however this was specifically tested after cycling. It would therefore be interesting for future research to examine whether selection on resilience occurs during different cycles of harsh drying and rewetting under laboratory circumstances.

Another point to make is that the samples were taken from the field at the end of April, before the summer drought treatment of 2017 had started. Based on this experiment we would expect to find more drought tolerant communities when sampling at the end of the summer treatment, since microbes adapt quickly to seasonal changes (Zeglin et al., 2013; Evans & Wallenstein, 2012). However, these seasonal adaptations might also act as a confounding factor and interfere with the legacy effects of historical treatment looked for in this research. For further research, it would be interesting to investigate how the behaviour of soils with a history of summer drought change during the growing season, thus when the drought treatment is active. As this research confirmed, microbial community shifts can occur after a single harsh drying and rewetting event (Fierer et al., 2003; Evans & Wallenstein, 2012). Therefore, the prediction would be that during one season a shift from type 2 to type 1 might occur, and that the drought treated microbial communities would experience this shift earlier than the communities from the control treatment.

(22)

21

When this experiment would be repeated it is worth paying attention to the following remarks. This study focussed on the evolution of responses of bacterial growth and respiration for the drying and rewetting cycles, however as mentioned before fungi also play an important role in soils. Specifically, since soils which are rich in fungal communities are better in keeping nutrients when subjected to cycles of D/RW, and soils with high SOM levels are suitable for fungi accommodation (Gordon et al., 2008). For further research, it would be interesting to investigate the response evolution of fungal growth for different perturbation cycles. This in order to assess the strength of the relationship between bacterial and fungal growth with a history of experimental summer drought. During sampling, some mineral matter was sampled from underneath the organic layer for the second control soil (C2). This resulted in a soil organic matter percentage of 55%, compared to SOM percentages of 70 to 85% for the other samples. This might have influenced the similarity of the replicate measurements. Moreover, an increased experiment size would lead to more available data for statistical analysis, and with that clear the road for more detailed analyses.

5. Conclusion

The soil studied in this experiment was an organic horizon, thus high in resources but also directly influenced by changes in temperature and precipitation. Microbial communities experience stress from cycles of (partial)drying and rewetting which works as a selective pressure for more drought tolerant microbial communities. This study showed that microbial communities which received historical summer drought for over 15 years shifted to being more resistant and resilient to drought. Harsh drying did result in a shift from a type 2 to a type 1 response. The soils which received a drought treatment did show significantly different behaviour in the first cycle of drying and rewetting, where the time lag after rewetting was shorter compared to the non-treated soils. Cycles of drying and rewetting, under laboratory conditions, did not result in selection for more resistant microbial communities, which raised the hypothesis that selection on resilience has occurred. The aim of this study was to increase the understanding of microbial responses upon changing climatic conditions. The first part of this study focussed on the field experiment effect on microbial communities. The results show that microbial communities can adapt to changes in climatic conditions even if those are only present annually for 3 months per year. The responses to the different cycles of drying and rewetting showed that already after one harsh drying event the soils shifted to a more drought enduring response type, which is type 1. Besides, both field drought treatment and laboratory drought perturbation resulted in increased carbon use efficiencies. The present study showed that soils subjected to (partial) drought in the field shifted towards higher resilience and resistance to drought, harsh drying and rewetting resulted in increased resilience, but not resistance. Both field drying and laboratory drying perturbation resulted in communities with higher carbon use efficiencies. With the prediction of increased intensity but specifically increased periodicity of precipitation, soil microbial communities from the organic horizon will presumably adapt and thus might mitigate the effect of climate change on ecosystems.

(23)

22

6. Acknowledgements

I would like to thank my supervisor dr. A. Tietema of the Institute of Biodiversity and Ecosystem Dynamics of the University of Amsterdam, who provided valuable comments and helped me keeping focus during the project.

Foremost, I would like to thank the research group on Microbial Ecology of Lund University who guided me through setting up and conducting a laboratory experiment. Special thanks to J. Rousk who supported me in doing my own research but guided me in the right direction when this was needed. In-depth feedback and valuable discussions helped me gaining insights into the world of microbial communities.

Besides my supervisors I would also like to thank H.P. Sterk for accompanying me to the experimental field site and providing a second opinion during my writing process. This helped me to focus on the bigger picture and not get caught up in details.

(24)

23

7. References

Bååth, E., Pettersson, M., & Söderberg, K. H. (2001). Adaptation of a rapid and economical microcentrifugation method to measure thymidine and leucine incorporation by soil bacteria. Soil Biology and Biochemistry, 33(11), 1571-1574.

Beier, C., Emmett, B., Gundersen, P., Tietema, A., Penuelas, J., Estiarte, M., ... & Williams, D. (2004). Novel approaches to study climate change effects on terrestrial ecosystems in the field: drought and passive night-time warming. Ecosystems, 7(6), 583-597.

Birch, H. F. (1958). The effect of soil drying on humus decomposition and nitrogen availability. Plant and soil, 10(1), 9-31. Borken, W., & Matzner, E. (2009). Reappraisal of

drying and wetting effects on C and N mineralization and fluxes in soils. Global

change biology, 15(4), 808-824.

Evans, S. E., & Wallenstein, M. D. (2012). Soil microbial community response to drying and rewetting stress: does historical precipitation regime matter? Biogeochemistry, 109(1-3), 101-116.

Fierer, N., & Schimel, J. P. (2002). Effects of drying–rewetting frequency on soil carbon and nitrogen transformations. Soil Biology and

Biochemistry, 34(6), 777-787.

Fierer, N., Schimel, J. P., & Holden, P. A. (2003). Influence of drying–rewetting frequency on soil bacterial community structure. Microbial

ecology, 45(1), 63-71.

Fraser, F. C., Corstanje, R., Deeks, L. K., Harris, J. A., Pawlett, M., Todman, L. C., ... & Ritz, K. (2016). On the origin of carbon dioxide released from rewetted soils. Soil Biology and Biochemistry, 101, 1-5.

Herron, P. M., Stark, J. M., Holt, C., Hooker, T., & Cardon, Z. G. (2009). Microbial growth efficiencies across a soil moisture gradient assessed using 13 C-acetic acid vapor and 15 N-ammonia gas. Soil Biology and Biochemistry, 41(6), 1262-1269.

Hueso, S., García, C., & Hernández, T. (2012). Severe drought conditions modify the microbial community structure, size and activity in amended and unamended soils. Soil

Biology and Biochemistry, 50, 167-173.

Göransson, H., Godbold, D. L., Jones, D. L., & Rousk, J. (2013). Bacterial growth and respiration responses upon rewetting dry forest soils: impact of drought-legacy. Soil Biology and

Biochemistry, 57, 477-486.

Gordon, H., Haygarth, P. M., & Bardgett, R. D. (2008). Drying and rewetting effects on soil microbial community composition and

nutrient leaching. Soil Biology and Biochemistry, 40(2), 302-311.

Iovieno, P., & Bååth, E. (2008). Effect of drying and rewetting on bacterial growth rates in soil. FEMS Microbiology Ecology, 65(3), 400-407. IPCC, 2013: Summary for Policymakers. In:

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, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Kim, D. G., Vargas, R., Bond-Lamberty, B., & Turetsky, M. R. (2012). Effects of soil rewetting and thawing on soil gas fluxes: a review of current literature and suggestions for future research. Biogeosciences, 9(7), 2459.

Lauber, C. L., Strickland, M. S., Bradford, M. A., & Fierer, N. (2008). The influence of soil properties on the structure of bacterial and fungal communities across land-use types. Soil Biology and Biochemistry, 40(9), 2407-2415.

Lipson, D. A., Schadt, C. W., & Schmidt, S. K. (2002). Changes in soil microbial community structure and function in an alpine dry meadow following spring snow melt.

Microbial ecology, 43(3), 307-314.

Meisner, A., Bååth, E., & Rousk, J. (2013). Microbial growth responses upon rewetting soil dried for four days or one year. Soil

Biology and Biochemistry, 66, 188-192.

Meisner, A., Leizeaga, A., Rousk, J., & Bååth, E. (2017). Partial drying accelerates bacterial growth recovery to rewetting. Soil Biology and

Biochemistry, 112, 269-276.

Meisner, A., Rousk, J., & Bååth, E. (2015). Prolonged drought changes the bacterial growth response to rewetting. Soil Biology

and Biochemistry, 88, 314-322.

Miller, A. E., Schimel, J. P., Meixner, T., Sickman, J. O., & Melack, J. M. (2005). Episodic rewetting enhances carbon and nitrogen release from chaparral soils. Soil Biology and Biochemistry, 37(12), 2195-2204.

Pulleman, M., & Tietema, A. (1999). Microbial C and N transformations during drying and rewetting of coniferous forest floor material. Soil Biology and Biochemistry, 31(2), 275-285.

Rath, K. M., Maheshwari, A., Bengtson, P., & Rousk, J. (2016). Comparative toxicities of salts on microbial processes in soil. Applied and

(25)

24

environmental microbiology, 82(7),

2012-2020.

Reichstein, M., Bahn, M., Ciais, P., Frank, D., Mahecha, M. D., Seneviratne, S. I., ... & Papale, D. (2013). Climate extremes and the carbon cycle. Nature, 500(7462), 287-295.

Rousk, J., Smith, A. R., & Jones, D. L. (2013). Investigating the long‐term legacy of drought and warming on the soil microbial community across five European shrubland ecosystems. Global change biology, 19(12), 3872-3884.

Schimel, D. S. (1995). Terrestrial ecosystems and the carbon cycle. Global change biology, 1(1), 77-91.

Tiemann, L. K., & Billings, S. A. (2011). Changes in variability of soil moisture alter microbial community C and N resource use. Soil Biology

and Biochemistry, 43(9), 1837-1847.

Xiang, S. R., Doyle, A., Holden, P. A., & Schimel, J. P. (2008). Drying and rewetting effects on C and N mineralization and microbial activity in surface and subsurface California grassland soils. Soil Biology and Biochemistry, 40(9), 2281-2289.

Zeglin, L. H., Bottomley, P. J., Jumpponen, A., Rice, C. W., Arango, M., Lindsley, A., ... & Myrold, D. D. (2013). Altered precipitation regime affects the function and composition of soil microbial communities on multiple time scales. Ecology, 94(10), 2334-2345.

Zwietering, M. H., Jongenburger, I., Rombouts, F. M., & Van't Riet, K. (1990). Modeling of the bacterial growth curve. Applied and environmental microbiology, 56(6),

(26)

25

8. Appendix

The raw-data is stored separately in .csv files with corresponding metadata files. Folder; denijs_2017_BSc_thesis_microbial_responses_data drying_rewetting_cycles_bacterialgrowth_data.csv drying_rewetting_cycles_bacterialgrowth_metadata.txt drying_rewetting_cycles_respiration_data.csv drying_rewetting_cycles_respiration_metadata.txt moisture_dependence_bacterialgrowth_data.csv moisture_dependence_bacterialgrowth_metadata.txt moisture_dependence_respiraiton_data.csv moisture_dependence_respiration_metadata.txt

8.1 Drying and rewetting time plan

(27)

26

8.2 Raw excel graphs

8.2.1 Initial moisture dependence data; raw graphs

Graph 1; initial moisture dependence, respiration rates

Graph 2; initial moisture dependence, bacterial growth

00 05 10 15 20 25 0,0 10,0 20,0 30,0 40,0 50,0 60,0 Re sp irat ion ra te (µ g CO2/g so il /h )

% of Water Holding Capacity

Initial moisture dependence; Respiration

C1 C2 C3 D1 D2 D3 0 100 200 300 400 500 600 0,0 10,0 20,0 30,0 40,0 50,0 60,0 Bacterial growth (p m o l Leu/g/h )

Initial moisture dependence; Bacterial Growth

C1 C2 C3 D1 D2 D3

(28)

27 8.2.2 Drying and rewetting data; raw graphs

Graph 3; drying and rewetting cycle 1, respiration rates

Graph 4; drying and rewetting cycle 1, bacterial growth

0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 Re sp irat ion ra te (µ g CO2/g/ h )

Time after rewetting (h)

D/RW 1 Respiration

C1 C2 C3 D1 D2 D3 0 100 200 300 400 500 600 700 0 10 20 30 40 50 60 70 80 Bac Growth (p m o l Leu/g/h )

D/RW 1 Bacterial growth

C1 C2 C3 D1 D2 D3

(29)

28

0 5 10 15 20 25 30 35 40 45 50 0 10 20 30 40 50 60 70 80 Re sp irat ion ra te (µ g CO2/g/ h )

D/RW 2 Respiration

C1 C2 C3 D1 D2 D3 0 50 100 150 200 250 300 0 10 20 30 40 50 60 70 80 Bac Growth (p m o l Leu/g/h )

D/RW 2 Bacterial growth

C1 C2 C3 D1 D2 D3

(30)

29

0 5 10 15 20 25 30 35 40 45 50 0 10 20 30 40 50 60 70 80 Re sp irat ion ra te (µ g CO2/g/ h )

D/RW 3 Respiration

C1 C2 C3 D1 D2 D3 0 50 100 150 200 250 300 0 10 20 30 40 50 60 70 80 Bac Growth (p m o l Leu/g/h )

D/RW 3 Bacterial growth

C1 C2 C3 D1 D2 D3

(31)

30

8.2.3 Final moisture dependence data; raw graphs

Graph 9; final moisture dependence, cycled soils, respiration rates

Graph 10; final moisture dependence, cycled soils, bacterial growth

00 02 04 06 08 10 12 14 0,0 10,0 20,0 30,0 40,0 50,0 60,0 Re sp irat ion ra te (µ g CO2/g so il /h )

Cycled soils: final moisture dependence; respiration

C1 C2 C3 D1 D2 D3 0 50 100 150 200 250 0,0 10,0 20,0 30,0 40,0 50,0 60,0 Bacterial growth (p m o l Leu/g/h )

Cycled soils: final moisture dependence; bacterial growth

C1 C2 C3 D1 D2 D3

The impact of historical summer droughts on the resistance and resilience of microbial communities to moisture stress