How extreme is an extreme climatic event to a subarctic peatland springtail community?

Soil Biology & Biochemistry 59 (2013) 16e24

Contents lists available at SciVerse ScienceDirect

Soil Biology & Biochemistry journal homepage: www.elsevier.com/locate/soilbio

How extreme is an extreme climatic event to a subarctic peatland springtail community? Eveline J. Krab a, *, Irene M. Van Schrojenstein Lantman a, Johannes H.C. Cornelissen a, Matty P. Berg b a b

Systems Ecology, Department of Ecological Science, Faculty of Earth and Life Sciences, VU University Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands Animal Ecology, Department of Ecological Science, Faculty of Earth and Life Sciences, VU University Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 November 2012 Received in revised form 19 December 2012 Accepted 21 December 2012 Available online 19 January 2013

Extreme climate events are increasing in frequency and duration and may directly impact belowground foodwebs and the activities of component soil organisms. The soil invertebrate community, which includes keystone decomposers, might respond to these newly induced soil microclimate conditions by shifts in density, species composition, spatial patterning and/or functional traits. To test if and how short-term extreme climatic conditions alter the structure, the vertical stratification and the community weighted trait means of the springtail (Collembola) community in sub-arctic peatbogs, we experimentally subjected Sphagnum peat cores in a field setting to factorial treatments of elevated temperature and episodically increased moisture content. The large precipitation peaks did not affect the springtail community, but an average soil temperature increase of 4
C halved its density in the shallower peat layers, mainly caused by the reduced dominance of Folsomia quadrioculata. A hypothesized net downward shift of the surface-dwelling springtail community, however, was not observed. We observed species-specific responses to warming but the overall community composition in subsequent organic layers was not significantly altered. Although the effects of an extreme warming event on density, species composition and vertical stratification pattern seemed subtle, functional trait analysis revealed directional community responses, i.e. an overall increase of soildwelling species due to warming, even though warming did not alter layer-specific community weighted trait means. We suggest that subtle changes in moisture conditions, due to increased evapotranspiration, have decreased typically surface-dwelling species relative to soil-dwelling species. The extent to which this directional change in the community is maintained after an extreme event, and its costs for the community’s resilience to multiple sequential extreme events will consequently determine its longer-term effects on the community and on ecosystem functioning. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Climate change Collembola Functional traits Extreme weather events Soil invertebrates Soil microclimate Warming Precipitation Vertical distribution Blanket bog

1. Introduction One of the major concerns with respect to global climate change is an expected increase in frequency and duration of extreme climate events (Hansen et al., 2012). Relatively short periods of extreme temperatures, drought or high precipitation can lead to changes in species’ physiological performance, relative abundances or even local-to-regional extinctions and altered distribution patterns (Easterling et al., 2000; Smith, 2011). Arctic regions are known to be particularly sensitive to climate change and extreme climatic events (Marchand et al., 2005; Post et al., 2009) which may impact on ecosystem processes such as soil organic matter breakdown

* Corresponding author. E-mail addresses: [email protected], [email protected] (E.J. Krab). 0038-0717/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.soilbio.2012.12.012

through their effects on the decomposer community. For example, experimentally applying a short episode of extreme winter warming in sub-arctic tundra caused strong shifts in its soil invertebrate community structure (Bokhorst et al., 2012). Microbivorous springtails and mites are a keystone group of soil invertebrates in arctic ecosystems (Woodin and Marquis, 1997). Although they only modestly contribute to carbon (C) turnover, their control over the biomass and activity of microbial decomposers (Lavelle, 1997; Hättenschwiler et al., 2005) can have a large impact on nitrogen (N) and C dynamics (Osler and Sommerkorn, 2007). The net effects of their activity on these dynamics are mainly dependent on: 1) their abundance; because low grazing pressure on fungi can stimulate microbial activity (Lussenhop, 1992) while high grazing pressure slows down microbial activity (Bardgett et al., 1993), 2) their identity; different species assemblages might be functionally different (Faber and

E.J. Krab et al. / Soil Biology & Biochemistry 59 (2013) 16e24

Verhoef, 1991) and 3) the quality of the substrate; breakdown of more N-rich litter results in net N mobilization while litter with a high (easily degradable) C content leads to immobilization of N (Berg, 2000). However, these factors vary considerably throughout most soils. In Arctic soils, strong short-scale vertical gradients of microclimate (temperature and humidity) and quality (in terms of C/N or lignin/N ratios) of decomposing litter and soil invertebrate abundance exist. Temperatures generally decline strongly down the soil profile, whereas deeper layers tend to be more humid. Shallow organic layers contain relatively easily degradable sugars, while in deeper organic layers lignin/N ratios tend to get higher and litter quality declines (Berg, 2000). These short-scale gradients in microclimate and substrate quality are generally considered to be the main responsible for soil invertebrate vertical distribution (Ponge, 2000; Berg and Bengtsson, 2007; Krab et al., 2010) as soil invertebrate abundances decline deeper down the soil profile (Ponge, 2000) and organic layers host typical species assemblages (Berg et al., 1998; Faber and Joosse, 1993; Krab et al., 2010). Global warming will increase summer air and soil temperature and also short periods of heavy rain in arctic regions (ACIA, 2005), which can soak peatland soils. Extreme events will alter the shortscale gradient in soil temperature and moisture across soil depth. These new microclimatic conditions potentially affect the activity and vertical distribution of soil invertebrates, which are known to be sensitive to changes in moisture and temperature (Lindberg et al., 2005; Huhta and Hanninen, 2001; Krab et al., 2010; Makkonen et al., 2011). The responses of soil invertebrates to these changes seem to be species-specific and dependent on their functional traits. For example soil-dwelling, euedaphic species and surface-dwelling, epigeic species, seem to respond stronger to alterations in microclimatic conditions than intermediate located hemiedaphic species or vertically broad-ranging species (Krab et al., 2010; Makkonen et al., 2011; Bokhorst et al., 2012). This implies that changes in soil microclimate due to climate change might lead to new soil invertebrate community structures, where both overall abundances and layer-specific species composition could change. Changes in community structure are in turn expected to impact processes to which soil invertebrates contribute, since functional roles of different soil invertebrates in C and N cycling vary vertically between decomposing layers (Faber, 1991; Kandeler et al., 1999), and between species (Heemsbergen et al., 2004; Chahartaghi et al., 2005). Climate warming has been shown to have potentially large effects on soil processes protruding down into the soil (Dorrepaal et al., 2009), and these observed effects have been shown to be strongly linked to increased activity of soil invertebrates deeper down the soil profile (Briones et al., 2007). However, to our knowledge, it has never been explicitly tested whether vertical distribution shifts of the soil invertebrate community occur due to changes in precipitation, temperature, or to both in combination, and whether such distribution shifts are community-wide or species-specific. Due to the interaction between warming and soil moisture content, peatland soil invertebrate community responses to temperature changes or alterations in moisture content could not be tested explicitly in previous studies (Shaver et al., 2006; Strack et al., 2009; Makkonen et al., 2011). Here, we present the first explicit test of such interactions through a unique factorial experiment in which we subjected Sphagnum peat cores to elevated temperatures and episodically increased moisture content in situ in a sub-arctic peatland and investigated springtail community response to changes in soil microclimate. We hypothesize that the warming treatment will cause a net downward shift of springtails (due to increased evaporation causing drought), while moisture addition

17

will cause a net upward shift (due to more beneficial moisture conditions in the top layers for soil-dwelling springtails). Further, we expect these shifts to differ between species, and to lead to altered layer-specific species assemblages. Finally, we will assess if species-specific traits that relate to microclimate, i.e. moisture preference, vertical stratification preference and body size, can be used to reveal mechanisms behind community responses. 2. Methods 2.1. Study site The experiment was conducted on a blanket bog on a slope near lake Torneträsk, in Abisko, Sweden (68
210 N, 18
490 E), at an elevation of 340e370 m above sea level. The average summer temperature is 11
C and there is an average annual precipitation of 300 mm (Abisko Scientific Research Station, Meteorological Station). This blanket bog is dominated by the moss Sphagnum fuscum (Schimp.) H. Klinggr., which grows intermingled with vascular plants (with a cover of about 25%) that mainly consists of Empetrum hermaphroditum Hagerup., Betula nana L., and Calamagrostis lapponica (Wahlenb.) Hartm. (Aerts et al., 2009). The experiment ran for 17 days, from the 28th of July until the 14th of August 2011, which is in the middle of the growing season. 2.2. Experimental setup The experiment followed a full factorial randomized block design with five blocks, each consisting of four treatments: a control, a water addition treatment, a warming treatment and a combined warming and water addition treatment. The treatments were executed in plots of 70  70 cm that each contained two transparent Perspex cylinders (Ø 12 cm, 20 cm long) that were inserted into the peat profile (the top flush with the surface) in the centre of the plot one day before the start of the experiment. The cylinders were placed at least 10 cm apart. Pilot studies had shown that these Perspex cylinders were necessary to prevent horizontal efflux of water in the precipitation treatments to the untreated surroundings, which acted as a sink for additional water. One of each pair of Perspex cylinders was used for sampling springtails while the other was used for logging temperature and moisture content, limiting disturbance to the soil fauna. Warming was obtained using Perspex tents. The tents were octagonal (Ø 53 cm, surface area w 2780 cm2) 22 cm high Perspex walls covered with a transparent plastic sheet, with 2 small openings (5 cm2) between the walls and the sheet for ventilation. The tents were placed in the centre of the 70  70 cm plots covering the two Perspex cylinders that were centred in the plots. The water addition treatment consisted of simulating a total additional daily precipitation of 22.5 mm (382.5 mm over the entire experiment), by steadily watering each plot with 3.7 l water three times a day using a watering can (intervals of approximately 6 h). We used water from lake Torneträsk to mimic extreme precipitation events (Makkonen et al. 2012). This amount of precipitation reflected realistic values for a very rainy day or three heavy rain showers in the area (Abisko Scientific Research Station, Meteorological Station). Since the tents were blocked off from precipitation, the amount of ambient rainfall was compensated for in these plots. 2.3. Microclimate measurements In four of the five blocks, plots had one soil moisture sensor (EC5 Soil Moisture Smart Sensor e S-SMC-M005) and one temperature sensor (12-Bit Temp Smart Sensor e S-TMB-M006) inserted vertically in the Sphagnum core contained by the Perspex cylinder; both

18

E.J. Krab et al. / Soil Biology & Biochemistry 59 (2013) 16e24

were connected to a data logger (HOBO Weather station e H21001). The temperature (
C) was measured at a depth of 4.5 cm and the volumetric water content (cm3 H2O cm3) across the depth of 3 to 9 cm. The loggers sampled soil temperature and moisture every 5 min. The loggers were inserted a day before starting the experiment, and recovered a day before the end of the experiment. During the experiment the tents increased average soil temperatures in the Perspex cylinders at 4.5 cm depth from an average of 11.5
C in the control plots to 15.5
C in the warmed plots (Fig. S1a). Water addition increased the average volumetric moisture content of the soil from 0.43 cm3 H2O cm3 to 0.50 cm3 H2O cm3 (Fig. S1b). In addition to the loggers, the temperature gradient was measured (to the nearest 0.1
C) at 1.5, 4.5, 7.5 and 10.5 cm depth, with a handheld soil thermometer, at five random days during the experiment. These soil depths correspond with the organic layers at which the springtails were recovered at time of harvest. Volumetric water content of the layers at time of harvest was calculated by weighing the Sphagnum core slices directly after harvest and after drying the cores for 48 h, at 70
C after Tullgren fauna extraction. 2.4. Harvest The cores were harvested after 17 days of climate treatment (3 h after applying the last moisture treatment). The Perspex cylinders were carefully removed from the peat profile and the contained Sphagnum cores were cut with a knife into four slices of organic layer of 3 cm each: (i) the top layer, consisting of living moss capitulum mixed with fresh moss litter, (ii) the second layer, which was made up of new and fragmented moss litter, (iii) the third layer consisting of more fragmented moss litter, and (iv) the oldest and deepest layer, comprising fragmented moss litter mixed with older peat. The slices were immediately sealed into plastic bags with sufficient air to prevent compaction, kept cooled (4
C), and transported in boxes to Amsterdam by plane for extraction within 48 h. Here the cores were stored at 4
C overnight whereafter springtails were extracted using a Tullgren Funnel fauna extractor (Van Straalen and Rijninks, 1982) and stored in 70% ethanol. Subsequently, the springtails were determined to species level using Fjellberg’s keys (1998, 2007) and counted. 2.5. Calculations and statistics The effects of warming, water addition and its combination on soil climatic conditions (logger data) were tested using a two-way ANOVA (there was no significant block effect). To test for effects of treatments on soil temperature and moisture gradient (manual measurement data), we used three-way ANCOVAs. ANCOVAs were used because conditions in the layers of a core are not independent from one another. Therefore we implemented layer as a covariate, fixed factors; warming and water addition and the random factor; block. Analyses for temperature data were conducted separately for measurements on different days. When the Levene’s tests showed that the error variance of the dependent variable was unequal between groups, we proceeded with ANCOVA if residual plots showed no consistent pattern in variance among treatments. To test if springtail community composition was affected by climate treatments (and thus if dissimilarities between communities living in cores undergoing different treatments were significant) we performed a permutation test for multifactorial multivariate analysis of variance (Two-way PERMANOVA, distance matrix: BrayeCurtis, no. of permutations: 999). The test statistic is a multivariate analogue to Fishers F-ratio and is calculated directly from a BrayeCurtis similarity matrix (Bray and Curtis, 1957). P

values are then obtained using permutations (Anderson, 2001). We did not include block effects in this analysis since the few observed block effects were subtle and not consistent (see Results). This test was carried out using the ‘vegan’ package in R (version 2.13.0). Changes in total springtail density and vertical distribution (density) due to our treatments were tested by three-way analysis of covariance (ANCOVA). The organic layer of the core (0e3 cm, 3e 6 cm, 6e9 cm, or 9e12 cm deep) was the covariate while presence or absence of warming and water addition and block were the two fixed- and single random factors, respectively. To test if vertical distribution shifts differed between species, individual three-way ANCOVAs on densities of the six most dominant species were performed. These species were: Folsomia quadrioculata (Tullberg, 1871), Friesea truncata Cassagnau, 1958, Ceratophysella denticulata (Bagnall, 1941), Protaphorura pseudovanderdrifti (Gisin, 1957), Parisotoma notabilis (Schäffer, 1896), and Desoria hiemalis (Schött, 1893) (See Table S2 for a complete species list). For the species that showed significant shifts in their density in the organic layers, post hoc Tukey tests following separate one-way ANOVAs were used to identify which treatment caused the observed significant effects. Where needed abundance data were transformed (log (nþ1)) to meet the assumptions of homogeneity of variance. We calculated BrayeCurtis similarity of the springtail communities in corresponding organic layers for different climate treatments. Similarities between springtail communities in the control treatment and the three climate treatments were calculated per treatment and per layer. To test if differences were significant we performed two-way PERMANOVAs using the ‘vegan’ package in R (version 2.13.0). Finally, to assess if the observed (layer-specific) changes in springtail community composition lead to an altered functional trait composition, we calculated the Community Weighted Means (CWM) of three functional traits: moisture preference, vertical stratification preference and maximum body size. For each trait we calculated CWMs for each of the three climate treatments and the control, and for all identified springtail species in the community. The trait values of each species were obtained from literature (Kuznetsova, 2003; Fjellberg, 1998, 2007), and springtail species’ moisture preference and vertical stratification were assessed by subdividing them into classes (Table S1). We followed Garnier et al. (2004) and Leps et al. (2006) to calculate the CWM for a whole community as: nj X

CWMj ¼

Ak;j  FTk;j

k¼1

where nj is the number of species sampled in community j, Ak,j is the relative abundance of species k in community j and FTk,j is the functional trait of interest of species k in community j. Differences in CWM between (layer-specific) communities were tested using a one-way ANCOVA. The organic layer of the core was the covariate while treatment was used as a fixed factor. We did not include block effects in this analysis as the few earlier observed block effects were consistent (see Results). A post hoc Tukey test following this analysis was used to identify which treatment dominated the observed significant effects. The PERMANOVA tests and statistical analyses on CWM were carried out using R (version 2.13.0) the other analyses were carried out using SPSS (version 19.0.). 3. Results 3.1. Abiotics Soil temperatures at 4.5 cm depth were substantially increased by the warming tents (F3,12 ¼ 186.0, P < 0.001) (Table 1),

E.J. Krab et al. / Soil Biology & Biochemistry 59 (2013) 16e24 Table 1 Summary of two-way ANOVAs performed on moisture and temperature logger data (n ¼ 4). Temperature data has been log transformed. Response variable Treatment

Warming Water addition Warming  water addition

Moisture content

Temperature

F

P

F

P

0.97 4.77 0.86

0.39 < 0.05 0.372

185.6 2.08 1.27

<0.001 0.175 0.372

Degrees of freedom (d.f) are 3.

which resulted in an average increase of 3.7  1.9
C and 4.6  2.3
C for the warming treatment and the combined water addition and warming treatment, respectively, without an interaction effect (Table 1). There was a strong diurnal cycle in soil temperatures; differences in soil temperature between treatments were stronger during daytime than at night (Fig. S1a). Manual temperature measurements also showed a strong effect of warming tents on soil temperature (F1,68 ranging from 135.4 to 190.4, P < 0.001) and there was generally no significant effect of water addition on soil temperature (Fig. 1, Table S2). Warming also increased the vertical soil temperature gradient considerably (Table S2). (F1,68 values for different days ranging from 10.2 to 35.5, P < 0.01). On three of the five measurement days water addition significantly increased layer specific soil temperatures (F1,68 values for 3 days ranging from 5.1 to 11.6, P < 0.05) (Fig. 1, Table S2). Soil moisture at a depth of 3 to 9 cm down the soil profile was significantly increased by water addition (F3,12 ¼ 4.77, P < 0.05, Fig. S1b). Water addition increased the average moisture content of the soil from 0.43 to 0.50 cm3 H2O cm3 (Fig. S1b). However, in gravimetrical measurements at the day of harvest this increase could not be detected. In addition, the soil moisture gradient was not significantly changed by the water addition treatment (Fig. 1, Table S2). Warming did not significantly affect total or layer specific soil moisture (Fig. 1, Table S2). 3.2. Springtail density and composition Total springtail density (730  93 ind. dm3) was significantly reduced by warming (504  46 ind. dm3) (F1,68 ¼ 7.24, P < 0.01) (Fig. 2, Table 2). BrayeCurtis similarities in species composition did not differ significantly from between treatments for whole cores (Table 3). This indicates that treatments did not significantly affect springtail species composition (relative species densities, as well as species presence or absence) at least when vertical stratification along the peat core is not taken into account.

a

19

3.3. Species vertical distribution Springtail density differed strongly between organic layers (F1,68 ¼ 60.9, P < 0.001). The top layers of the control hosted a significantly higher springtail density than its deeper layers (One-way ANOVA, F3,16 ¼ 6.65, P < 0.01) (Fig. 2a). Average density in the top layer of the control was 1410  666 ind. dm3 whereas densities in the deeper layers were 625  219, 528  272 and 358  158 ind. dm3, respectively. This pattern changed significantly owing to warming (F1,68 ¼ 4.41, P < 0.05) (Table 2), after which the difference between the density in the top and the deeper organic layers was less pronounced (Fig. 2.) Total springtail density in the top layer dropped by 49%, from 1410  666 in controls to 719  329 ind. dm3 in the warming treatment. Most individual springtail species showed a significant vertical stratification pattern (Fig. 2, Table 2), with typical surface dwelling species, such as F. quadioculata, F. truncata and C. denticulata mostly abundant at the top layers and soil-dwelling species, such as P. pseudovanderdrifti mostly occurring in deeper layers. The only species that significantly changed their vertical stratification due to climate manipulation were F. quadrioculata and P. pseudovanderdrifiti. The vertical distribution of F. quadrioculata responded to temperature increase, as its density in the top layer declined from 1003  695 to 329  229 ind. dm3 (Fig. 2). There was no significant interaction between the ‘warming and water addition’ treatment and organic layer (Table 2), suggesting that ‘warming and water addition’ did not alter vertical distribution shifts compared to the warming only. Although densities of P. pseudovanderdrifti were very variable between layers, no overall significant difference in density between layers was observed, with the exception of warming treatment (Table 2). However, there was a significant interaction between warming and organic layer, indicating that its vertical distribution pattern shifted due to temperature increase. The other dominant species F. truncata, C. denticulata, P. notabilis and D. hiemalis did not show a significant response in their vertical distribution to our climate treatments. Although some species changed their vertical stratification pattern, these shifts did not lead to significant shifts in layer-specific community structure, as the two-way PERMANOVA tests showed no significant organic layer-specific springtail community shifts (Table 3).

3.4. Community weighted mean (CWM) trait values CWMs differed significantly between organic layers for all three traits (Fig. 3, Table S3). Vertical stratification preference and

b

Fig. 1. Temperature (a) and moisture (b) gradients down the peat profile for three climate treatments and a control (CC) (n ¼ 5). Climate treatments are water addition (CM), warming (TC) and the combination of warming and water addition (TM). Grey scales represent different layers down the peat profile. Temperature measurements are averages of five days and moisture gradients are those measured at time of harvest.

20

E.J. Krab et al. / Soil Biology & Biochemistry 59 (2013) 16e24

Fig. 2. Average density (ind. dm3) per peat layer of: (a) all springtail species, and six dominant species (beg) climate treatments are water addition (CM), warming (TC) and the combination of warming and water addition (TM). CC denotes the control. The order of the 3 bars and the grey scales per treatment represent the position of the different organic layers at time of harvest. Significance levels are P < 0.05. Bars that share the same letter are not significantly different from one another (Tukey tests). Within-treatment differences in springtail densities of different peat layers are not separately indicated if not mentioned in the ‘Results’ text. Error bars are standard errors (n ¼ 5).

maximum body size showed significant shifts due to our treatments. The CWM of vertical stratification preference showed a decrease in surface-dwelling species in the warmed community relative to control (Fig. 3, Table S3). The CWM for maximum body size increased with warming and in water addition  warming relative to the control. The average moisture preference of communities did not change. None of the CWMs traits showed a significant

interaction between layer and treatment, indicating that there were no significant layer-specific shifts in CWMs (Table S3). 4. Discussion Our results demonstrate that an average soil temperature increase of about 4
C significantly reduced springtail densities in

E.J. Krab et al. / Soil Biology & Biochemistry 59 (2013) 16e24

21

Table 2 Summary of the three-way ANCOVAs on springtail density (n ¼ 5). Fixed factors are warming and water addition, organic layers are covariates and block effect a random factor. Effects of climate treatments are tested by ‘warming’, ‘water addition’ and ‘warming  water addition’, the effect of organic layer/depth on springtail density is tested by ‘layer’. Significant effects of climate treatments on the vertical distribution of soil fauna are tested by climate treatment interactions with ‘layer’. Species

Main effects Warming

Folsomia quadrioculata Friesea truncata Ceratophysella denticulata Protaphorura pseudovanderd. Parisotoma notabilis Desoria hiemalis spp. Total

Interactions Water addition

Layers

Warming  water add.

Warming  layers

Water add. Layers

Warming  water add.  layers

F

P

F

P

F

P

F

P

F

P

F

P

F

P

9.87 0.00 3.88 3.86 0.45 0.95 7.24

<0.01 0.951 0.053 0.053 0.506 0.333 <0.01

0.55 0.65 3.61 3.41 2.87 1.11 1.84

0.46 0.424 0.062 0.069 0.095 0.295 0.18

19.00 120.6 23.42 0.683 11.02 0.826 60.85

<0.001 <0.001 <0.001 0.412 <0.01 0.367 <0.001

3.12 2.08 2.30 3.59 0.47 0.33 3.34

0.082 0.153 0.134 0.063 0.497 0.570 0.072

5.85 0.16 1.97 3.97 0.80 0.11 4.41

<0.05 0.695 0.165 <0.05 0.374 0.744 <0.05

0.04 0.96 2.30 3.41 1.24 2.31 0.63

0.852 0.332 0.134 0.069 0.269 0.133 0.429

1.00 2.04 1.20 2.51 0.32 0.75 1.31

0.321 0.158 0.277 0.118 0.576 0.390 0.257

Degrees of freedom are 1 for each main effect or interaction, error df are 68.

a northern peatland and by doing so also changed their vertical stratification pattern. However, this change did not result in a net downward shift of the community. It also did not significantly alter community structure since there were only minor shifts in the relative springtail species abundance. Despite the lack of community composition response based on taxonomy, the average community-weighted trait values (CWM) for vertical stratification preference changed towards more soil-dwelling species in the warmed treatment relative to the control, although we did not observe an organic layer-specific response. Surprisingly a very strong increase in precipitation only altered soil moisture conditions marginally, and did not significantly affect springtail densities, their vertical stratification pattern, the community structure or trait CWMs. However, before we can draw justifiable ecological conclusions from these results we will consider the strengths and drawbacks of our set-up and discuss our results in more detail. 4.1. An extreme precipitation event does not affect springtail community structure Although soil invertebrates are known to be very sensitive to changes in moisture conditions (O’Lear and Blair, 1999; Huhta and Hanninen, 2001; Lindberg et al., 2005) a strong simulated increase in precipitation did neither affect springtail densities, nor their vertical stratification pattern, the community structure or trait CWMs. Subsequently, we did not observe the hypothesized upward shift of springtails due to an increase in soil moisture. Surprisingly, soil moisture increased only slightly due to our simulated precipitation regime, and did not lead to changes in soil moisture across layers. To mimic precipitation events realistically with experimental water addition can be a challenge due to horizontal water runoff (Beier et al., 2012), but in our experiment the Perspex cylinders prevented horizontal efflux of water directly after watering. We believe that the created realistic moisture conditions following simulated heavy rain are due to vertical drainage to deeper layers. The actual physical impact of the falling water drops, potentially flushing small decomposers like testate amoeba out of the peat core (Tsyganov et al., 2013), did not seem

to affect the springtails. Also, peat core atmospheres are probably saturated or near-saturated with water vapour. Springtails might be more sensitive to this water vapour rather than to the added liquid water, and react to thresholds of moisture rather than gradients. This would be in contrast to the effect of temperature, which is more progressive. 4.2. Warming affects springtail density, community trait means but not community composition Our warming treatments increased temperatures considerably across the entire peat profile (on average 4
C over the first 12 cm), which would not be an unrealistic temperature increase, even on the longer term, and some proposed future climate scenarios (ACIA, 2005). Warming declined springtail density in the top organic layer, but did not increase its density in deeper organic layers, implying that there was no net downward shift of the community due to warming. Climate warming therefore potentially has a negative effect on surface-dwelling species. Although these species are generally known to be less sensitive to extreme warming than deeper living species (Van Dooremalen et al. in press) their relative abundance decreased. An explanation could be that surfacedwelling species, which dominate in shallow peat layers, experienced more extreme conditions than soil-dwelling species, since climate warming mainly impacts on soil temperature in shallow layers. Further, these deeper peat layers might be species-saturated or restricted in resources to allow for a downward shift of surfacedwelling species. Alternatively surface-living species might be unable to sustain specific environmental conditions (such three dimensionality, CO2 concentration and food resources) in deeper layers. We observed species-specific responses. Two species, F. quadrioculata and P. pseudovanderdrifti, changed their vertical stratification pattern in response to warming, where the other dominant species did not. The dominant species F. quadrioculata showed a warming-induced decline in abundance in the top layer probably due to the relatively strong temperature increase. Folsomia quadrioculata is considered to be relatively robust to changes in

Table 3 Two-way PERMANOVA analyses (distance matrix: BrayeCurtis, no of permutations: 999) carried out for each peat layer separately (0e3, 3e6, 6e9 and 9e12 cm) resulting in four PERMANOVA analyses. Fixed factors were warming, water addition, and warming  water addition. Treatment

Total F

P

F

P

F

P

F

P

F

P

Warming Water addition Warming  water addition

1.70 0.80 1.51

0.173 0.410 0.209

2.17 0.78 0.89

0.099 0.478 0.396

1.4 2.45 0.53

0.234 0.083 0.655

0.38 0.97 0.84

0.79 0.359 0.434

0.26 0.62 0.53

0.885 0.565 0.624

Degrees of freedom (d.f.) were 1 in all analyses.

0e3 cm

3e6 cm

6e9 cm

9e12 cm

22

E.J. Krab et al. / Soil Biology & Biochemistry 59 (2013) 16e24

strong warming i.e. an increase in CO2 concentration due to increased microbial activity in surface layers. In addition, Protaphorura pseudovanderdrifti shifted upwards towards higher temperatures rather than downwards towards the lower ones, even though it is known to be sensitive to microclimate and to prefer deeper, colder and wetter, peat layers over the warmer, drier and less temperature buffered surface layers (Krab et al., 2010). Other species, such as C. denticulata and P. notabilis showed similar (although not significant) shifts towards the top layer. An explanation could be that the decrease in abundance of the dominant F. quadrioculata from the top layers reduced the competition for resources in that layer, suggesting that interspecific competition might also be a strong driver for springtail vertical distribution. Unexpectedly, the specific responses of F. quadrioculata and P. pseudovanderdrifti species did not result in a significantly altered community structure, as there were only minor shifts in the relative abundances of springtail species. However, warming did cause a shift in the trait CWM of the springtails community in peat columns; we observed a decrease in the average vertical stratification preference and an increase in the average body size of the community due to warming. Trait CWMs of specific organic layers did not change due to warming. Seemingly unimportant minor shifts in relative abundances can result in relevant shifts in trait means. Therefore, when climate-driven community changes are only assessed using taxonomic information, the consequences of such shifts can be underestimated. The observed changes due to warming are potentially related to warming induced drought. The combined warming and moisture treatment did not change vertical springtail distribution patterns or density relative to the control, whereas the warming treatment did. Therefore the significant effect of the warming seemed to be linked to moisture limitation. Furthermore the shift in CWM for vertical distribution preference was observed for the warming treatment but not for the warming cum water addition treatment. These observations indeed suggest warming-induced drought to be the driving factor for springtail responses, even though our microclimatic data only showed a non-significant trend for warmed peat cores to be drier (Fig. 1) which was most pronounced in the deeper organic layer. We therefore carefully suggest that warmingincreased drought indeed is an important driving factor for springtail responses, rather than the temperature increase itself. 4.3. Springtail responses to microclimate related to species traits Fig. 3. Community weighted mean trait values (CWM) for springtail communities undergoing four extreme climate treatments: control, moisture addition, warming, and moisture addition and warming. CWMs are calculated for organic layer specific communities, where L represents the litter layer (0e3 cm deep), F is the fragmented litter layer (3e6 cm deep), H is the humus layer (6e9 cm deep) and H2 is the deeper humus layer (9e12 cm deep). The traits for which CWMs are calculated are: a) moisture preference, b) vertical distribution preference and c) maximum body size. For moisture preference a higher value indicates a more hygrophilous community, and for vertical distribution preference a higher value represents a more surface-dwelling species in the community. Maximum body size is expressed in mm. Springtail communities with a ‘*’ have a significantly different CWM than the control treatment. Error bars are standard deviations (n ¼ 5).

(micro)climate (Birkemoe and Sømme, 1998; Krab et al., 2010), although it is also known to be moderately sensitive to drought (Hertzberg and Leinaas, 1998). It showed a positive response to an extreme winter warming event in a nearby sub-arctic heathland (Bokhorst et al., 2012) and no significant reductions due to experimental warming in sub-arctic heathland (Sjursen et al., 2005; Makkonen et al., 2011), therefore its response to our warming treatment was somewhat unexpected. Potentially this species might have been negatively affected by peat-specific side effects of

Soil invertebrates differ in traits that underlie their sensitivity to microclimate and substrate quality (Krab et al., 2010; Makkonen et al., 2011). Therefore we expected species to respond differently to the climate manipulation treatments, and to observe changes in (layer-specific) community composition. We did observe differences between species as some were responsive to changes in microclimate and others were not. CWMs of body size and vertical stratification preferences changed significantly due to warming, but changed in the opposite direction than could be expected from the results of previous warming experiments. Vertical stratification preference shifted towards more soil-dwelling species in the warming treatment, whereas Makkonen et al. (2011) showed a relative increase in surface-dwelling species due to the effects of experimental warming on soil moisture conditions. We suggest that the strong warming in the top layer caused the loss of more surface-dwelling species. Although there was no significant layerspecific effect of warming for CWMs, there was a decline in springtail density in the top layer of our warming treatment. This decline has led to a relative decrease in epigeic species relative to euedaphic species over the first 12 cm of the peat profile.

E.J. Krab et al. / Soil Biology & Biochemistry 59 (2013) 16e24

We also observed an increase in average maximum body size due to warming. This result does not concur with that of vertical stratification preference, as euedaphic species are typically smaller than epigeic species (Makkonen et al., 2011; Bokhorst et al., 2012) although Kuznetsova (2003) considered some smaller species in our Sphagnum community to be hemi-edaphic rather than euedaphic species (Table S1). Using trait values for maximum body size from the literature instead of actual body size might have caused this discrepancy, since body size is a plastic trait and can vary within species considerably in different habitats (Ulrich and Fiera, 2010) and between different life stages. Therefore we cannot exclude the possibility that the actual body sizes did not shift in response to warming in our experiment and cannot draw any firm conclusions from these results. We therefore suggest for future trait analyses to calculate CWM body size from the observed body size distribution in the community. 4.4. Conclusions Climate manipulation experiments quite regularly show relatively little response of soil invertebrate community structure to elevated temperature (Coulson et al., 1996; Sohlenius and Bostrom, 1999; Sjursen et al., 2005). However, most of those studies did not take shifts in vertical distribution and trait means into account. In the few studies where CWM trait values were incorporated in the analysis, communities that showed a lack in response based on species (relative) abundances did show significant shifts in trait CWMs (Bokhorst et al., 2012). Our study confirms these observations, since there was no effect of extreme precipitation and warming on the species composition, but warming did affect density and CWMs of the springtail community, and might potentially affect other groups of decomposers as well. These shifts can cause changes in ecosystem functioning as traits involved in the response of a community can also be linked to those causing the effects on ecosystem processes (Suding et al., 2008). Epigeic species, for example, graze on different fungal groups than euedaphic species, since not only their position in the soil but also the morphology of their mouthparts differs considerably (Fjellberg, 1998, 2007; unpublished data). By grazing on different fungal groups invertebrates can significantly alter fungal communities (Crowther and A’Bear, 2012) and their activity (Bardgett et al., 1993). Reversely, a response in fungal activity to the climate treatments, could impact on the springtail population size and location within the soil profile since fungi are considered to be the main food source for springtails. By disentangling warming and moisture effects, incorporating vertical stratification patterning and by assessing changes in community composition by means of trait CWMs, we have taken a significant step towards unravelling the mechanisms behind soil invertebrate responses to an extreme climatic event. However, a next critical step in researching the effects of climate extreme events on the decomposer community and its longer-term effects on carbon cycling, will be to study its resilience to multiple sequential extreme climate events. In a world where extreme climatic events will rather be rule than exception, a species’ plasticity in relation to changing microclimatic conditions might well be the key to its future success (Miner et al., 2005; Berg and Ellers, 2010). Acknowledgements We would like to thank the Abisko Scientific Research Station, Sweden, and several of its staff, for providing research facilities and hospitality. We would especially like to thank Dylan Jones for lending us Perspex warming tents. This study was made possible by funding of NWO (Netherlands Organization for Scientific Research),

23

specifically its International Polar Year program (ENVISNAR project) through grant 851.40.060. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.soilbio.2012.12.012. References ACIA, 2005. Arctic Climate Impact Assessment. Cambridge University Press, New York. Aerts, R., Callaghan, T.V., Dorrepaal, E., van Logtestijn, R.S.P., Cornelissen, J.H.C., 2009. Seasonal climate manipulations result in species-specific changes in leaf nutrient levels and isotopic composition in a sub-arctic bog. Functional Ecology 23, 680e688. Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance. Austral Ecology 26, 32e46. Bardgett, R.D., Whittaker, J.B., Frankland, J.C., 1993. The effect of collembolan grazing on fungal activity in differently managed upland pasturesea microcosm study. Biology and Fertility of Soils 16, 255e262. Beier, C., Beierkuhnlein, C., Wohlgemuth, T., Penuelas, J., Emmett, B., Korner, C., de Boeck, H., Christensen, J.H., Leuzinger, S., Janssens, I.A., Hansen, K., 2012. Precipitation manipulation experiments e challenges and recommendations for the future. Ecology Letters 15, 899e911. Berg, B., 2000. Litter decomposition and organic matter turnover in northern forest soils. Forest Ecology and Management 133, 13e22. Berg, M.P., Bengtsson, J., 2007. Temporal and spatial variability in soil food web structure. Oikos 116, 1789e1804. Berg, M.P., Ellers, J., 2010. Trait plasticity in species interactions: a driving force of community dynamics. Evolutionary Ecology 24, 617e629. Berg, M.P., Kniese, J.P., Bedaux, J.J.M., Verhoef, H.A., 1998. Dynamics and stratification of functional groups of micro- and mesoarthropods in the organic layer of a Scots pine forest. Biology and Fertility of Soils 26, 268e284. Birkemoe, T., Sømme, L., 1998. Population dynamics of two collembolan species in Arctic tundra. Pedobiologia 42, 131e145. Bokhorst, S., Phoenix, G.K., Bjerke, J.W., Callaghan, T.V., Huyer-Brugman, F., Berg, M.P., 2012. Extreme winter warming events more negatively impact small rather than large soil fauna: shift in community composition explained by traits not taxa. Global Change Biology 18, 1152e1162. Bray, J.R., Curtis, J.T., 1957. An ordination of the upland forest communities of Southern Wisconsin. Ecological Monographs 27, 326e349. Briones, M.J.I., Ostle, N.J., Garnett, M.H., 2007. Invertebrates increase the sensitivity of non-labile soil carbon to climate change. Soil Biology & Biochemistry 39, 816e818. Chahartaghi, M., Langel, R., Scheu, S., Ruess, L., 2005. Feeding guilds in Collembola based on nitrogen stable isotope ratios. Soil Biology & Biochemistry 37, 1718e 1725. Coulson, S.J., Hodkinson, I.D., Webb, N.R., Block, W., Bale, J.S., Strathdee, A.T., Worland, M.R., Wooley, C., 1996. Effects of experimental temperature elevation on high-arctic soil microarthropod populations. Polar Biology 16, 147e153. Crowther, T.W., A’Bear, A.D., 2012. Impacts of grazing soil fauna on decomposer fungi are species-specific and density-dependent. Fungal Ecology 5, 277e281. Dorrepaal, E., Toet, S., van Logtestijn, R.S.P., Swart, E., van de Weg, M.J., Callaghan, T.V., Aerts, R., 2009. Carbon respiration from subsurface peat accelerated by climate warming in the subarctic. Nature 460, 616e679. Easterling, D.R., Meehl, G.A., Parmesan, C., Changnon, S.A., Karl, T.R., Mearns, L.O., 2000. Climate extremes: observations, modeling, and impacts. Science 289, 2068e2074. Faber, J.H., 1991. Functional classification of soil faunaea new approach. Oikos 62, 110e117. Faber, J.H., Joosse, E.N.G., 1993. Vertical distribution of collembola in a Pinus-nigra organic soil. Pedobiologia 37, 336e350. Faber, J.H., Verhoef, H.A., 1991. Functional differences between closely-related soil arthropods with respect to decomposition processes in the presence or absence of Pine tree roots. Soil Biology & Biochemistry 23, 15e23. Fjellberg, A., 1998. The Collembola of Fennoscandia and Denmark, Part I: Poduramorpha. Brill, Leiden. Fjellberg, A., 2007. The Collembola of Fennoscandia and Denmark. Part II: Entomobryomorpha and Symphypleona. Brill, Leiden. Garnier, E., Cortez, J., Billes, G., Navas, M.L., Roumet, C., Debussche, M., Laurent, G., Blanchard, A., Aubry, D., Bellmann, A., Neill, C., Toussaint, J.P., 2004. Plant functional markers capture ecosystem properties during secondary succession. Ecology 85, 2630e2637. Hansen, J., Satoa, M., Ruedy, R., 2012. Perception of climate change. Proceedings of the National Academy of Sciences of the USA 109, 14726e14727. Hättenschwiler, S., Tiunov, A.V., Scheu, S., 2005. Biodiversity and litter decomposition interrestrial ecosystems. Annual Review of Ecology Evolution and Systematics 36, 191e218. Heemsbergen, D.A., Berg, M.P., Loreau, M., van Haj, J.R., Faber, J.H., Verhoef, H.A., 2004. Biodiversity effects on soil processes explained by interspecific functional dissimilarity. Science 306, 1019e1020.

24

E.J. Krab et al. / Soil Biology & Biochemistry 59 (2013) 16e24

Hertzberg, K., Leinaas, H.P., 1998. Drought stress as a mortality factor in two pairs of sympatric species of Collembola at Spitsbergen, Svalbard. Polar Biology 19, 302e306. Huhta, V., Hanninen, S.M., 2001. Effects of temperature and moisture fluctuations on an experimental soil microarthropod community. Pedobiologia 45, 279e286. Kandeler, E., Kampichler, C., Joergensen, R.G., Molter, K., 1999. Effects of mesofauna in a spruce forest on soil microbial communities and N cycling in field mesocosms. Soil Biology & Biochemistry 3, 1783e1792. Krab, E.J., Oorsprong, H., Berg, M.P., Cornelissen, J.H.C., 2010. Turning northern peatlands upside down: disentangling microclimate and substrate quality effects on vertical distribution of Collembola. Functional Ecology 24, 1362e1369. Kuznetsova, N.A., 2003. Humidity and distribution of springtails. Entomological Review 83, 230e238. Lavelle, P., 1997. Faunal activities and soil processes: adaptive strategies that determine ecosystem function. Advances in Ecological Research 27, 93e132. Leps, J., de Bello, F., Lavorel, S., Berman, S., 2006. Quantifying and interpreting functional diversity of natural communities: practical considerations matter. Preslia 78, 481e501. Lindberg, N., Bengtsson, J.B., Persson, T., 2005. Effects of experimental irrigation and drought on the composition and diversity of soil fauna in a coniferous stand. Journal of Applied Ecology 39, 924e936. Lussenhop, J., 1992. Mechanisms of microarthropod microbial interactions in soil. Advances in Ecological Research 23, 1e33. Makkonen, M., Berg, M.P., van Hal, J.R., Callaghan, T.V., Press, M.C., Aerts, R., 2011. Traits explain the responses of a sub-arctic Collembola community to climate manipulation. Soil Biology & Biochemistry 43, 377e384. Makkonen, M., Berg, M.P., Van Logtestijn, R.S.P., Van Hal, J.R., Aerts, R., 2012. Do physical plant litter traits explain non-additivity in litter mixtures? A test of the improved microenvironmental conditions theory. Oikos. http://dx.doi.org/ 10.1111/j.1600-0706.2012.20750.x. Marchand, F.L., Mertens, S., Kockelbergh, F., Beyens, L., Nijs, I., 2005. Performance of High Arctic tundra plants improved during but deteriorated after exposure to a simulated extreme temperature event. Global Change Biology 11, 2078e2089. Miner, B.G., Sultan, S.E., Morgan, S.G., Padilla, D.K., Relyea, R.A., 2005. Ecological consequences of phenotypic plasticity. Trends in Ecology & Evolution 20, 685e692. O’Lear, H.A., Blair, J.M., 1999. Responses of soil microarthropods to changes in soil water availability in tallgrass prairie. Biology and Fertility of Soils 29, 207e217. Osler, G.H.R., Sommerkorn, M., 2007. Toward a complete soil C and N cycle: incorporating the soil fauna. Ecology 88, 1611e1621. Ponge, J.F., 2000. Vertical distribution of Collembola (Hexapoda) and their food resources in organic horizons of beech forests. Biology and Fertility of Soils 32, 508e522.

Post, E., Forchhammer, M.C., Bret-Harte, M.S., Callaghan, T.V., Christensen, T.R., Elberling, B., Fox, A.D., Gilg, O., Hik, D.S., Hoye, T.T., Ims, R.A., Jeppesen, E., Klein, D.R., Madsen, J., McGuire, A.D., Rysgaard, S., Schindler, D.E., Stirling, I., Tamstorf, M.P., Tyler, N.J.C., van der Wal, R., Welker, J., Wookey, P.A., Schmidt, N.M., Aastrup, P., 2009. Ecological dynamics across the Arctic associated with recent climate change. Science 325, 1355e1358. Shaver, G.R., Giblin, A.E., Nadelhoffer, K.J., Thieler, K.K., Downs, M.R., Laundre, J.A., Rastetter, E.B., 2006. Carbon turnover in Alaskan tundra soils: effects of organic matter quality, temperature, moisture and fertilizer. Journal of Ecology 94, 740e 753. Sjursen, H., Michelsen, A., Jonasson, S., 2005. Effects of long-term soil warming and fertilisation on microarthropod abundances in three sub-arctic ecosystems. Applied Soil Ecology 30, 148e161. Smith, M.D., 2011. The ecological role of climate extremes: current understanding and future prospects. Journal of Ecology 99, 651e655. Sohlenius, B., Bostrom, S., 1999. Effects of climate change on soil factors and metazoan microfauna (nematodes, tardigrades and rotifers) in a Swedish tundra soil e a soil transplantation experiment. Applied Soil Ecology 12, 113e 128. Strack, M., Waddington, J.M., Lucchese, M.C., Cagampan, J.P., 2009. Moisture controls on CO2 exchange in a Sphagnum-dominated peatland: results from an extreme drought field experiment. Ecohydrology 2, 454e461. Suding, K.N., Lavorel, S., Chapin, F.S., Cornelissen, J.H.C., Diaz, S., Garnier, E., Goldberg, D., Hooper, D.U., Jackson, S.T., Navas, M.L., 2008. Scaling environmental change through the community-level: a trait-based response-and-effect framework for plants. Global Change Biology 14, 1125e1140. Tsyganov, A.N., Keuper, F., Aerts, R., Beyens, L., 2013. Flourish or flush: effects of simulated extreme rainfall events on Sphagnum-dwelling testate amoebae in a subarctic bog (Abisko, Sweden). Microbial Ecology 65, 101e110. Ulrich, W., Fiera, C., 2010. Environmental correlates of body size distributions of European springtails (Hexapoda: Collembola). Global Ecology and Biogeography 19, 905e915. Van Dooremalen, C., Berg, M.P., Ellers, J. Acclimation responses to temperature vary with vertical stratification: implications for vulnerability of soildwelling species to extreme temperature events. Global Change Biology, in press. Van Straalen, N.M., Rijninks, P.C., 1982. The efficiency of Tullgren apparatus with respect to interpreting seasonal changes in age structure of soil arthropod populations. Pedobiologia 24, 197e209. Woodin, S.J., Marquis, M., 1997. Ecology of Arctic Environments. Blackwell.