Electronic Scientific Paper Archive

Effects of open-top passive warming chambers on soil respiration in the semi-arid steppe to taiga forest transition zone in Northern Mongolia

Anarmaa Sharkhuu • Alain F. Plante • Orsoo Enkhmandal • Brenda B. Casper • Brent R. Helliker • Bazartseren Boldgiv • Peter S. Petraitis

Abstract

The response of soil respiration to warming has been poorly studied in regions at higher latitude with low precipitation. We manipulated air temperature, soil temperature and soil moisture using passive, open-top chambers (OTCs) in three different ecosystem settings in close proximity (boreal forest, riparian area, and semi-arid steppe) to investigate how environmental factors would affect soil respiration in these different ecosystems, anticipating that soil respiration would increase in response to the chamber treatment. The results indicated that OTCs significantly increased air and soil temperature in areas with open canopy and short-statured vegetation (i.e., steppe areas) but not in forest. OTCs also affected soil moisture, but the direction of change depended on the ecosystem, and the magnitude of change was highly variable. Generally, OTCs did not affect soil respiration in steppe and riparian areas. Although soil respiration was slightly greater in OTCs placed in the forest, the difference was not statistically significant. Analyses of relationships between soil respiration and environmental variables suggested that different factors controlled soil respiration in the different ecosystems. Competing effects analysis using a model selection approach and regression analyses (e.g., Q10) demonstrated that soil respiration in the forest was more sensitive to warming, while soil respiration in the steppe was more sensitive to soil moisture. The differing responses and controlling factors among these neighboring forest, riparia and steppe ecosystems in Northern Mongolia highlight the importance of taking into account potential biome shifts in C cycling modeling to generate more accurate predictions of landscape-scale responses to anticipated climate change.

Keyword: Passive open-top chamber,Warming,Soil respiration,Taiga Forest,Steppe,Mongolia

Introduction

Soil respiration plays an important role in terrestrial ecosystem carbon cycling. Globally, the flux of C to the atmosphere from terrestrial respiration is 6–7 times larger than current anthropogenic emissions (*60 vs * 9 Pg C year-1, IPCC 2007). As a major component of ecosystem respiration, soil respiration accounted for 71 % of terrestrial ecosystem respiration in a mixed hardwood (Curtis et al. 2005) and 52 % of ecosystem respiration in an alpine grassland (Zhang et al. 2009), and was positively correlated with litterfall amount in forests (Schlesinger 1977) and with net primary productivity in grasslands (Raich and Schlesinger 1992; Raich and Tufekcioglu 2000). Soil respiration measurement may therefore represent a good proxy for ecosystem carbon cycling rates. On a global scale, soil respiration increases with increasing air temperature (Bond-Lamberty and Thomson 2010), and thus potentially generates a positive feedback with warming (Heimann and Reichstein 2008). However, responses of carbon cycling to climate change varied drastically among different modeling simulations (Heimann and Reichstein 2008) and among experimental warming studies. Some field studies using experimental warming have shown that soil respiration increases (Biasi et al. 2008; Wan et al. 2005), while other studies have shown that it decreases (Liu et al. 2009) or does not change (Wan et al. 2007). These apparently contradictory results may be due to differing antecedent conditions among the studies, or due to confounding effects of altered environmental conditions generated by the warming apparatus used in the different studies. For instance, soil respiration in a warming chamber may respond positively to increased temperature, but may also respond negatively due to decreased moisture associated with rainfall interception by the chamber.
These varying results highlight that effects of climate change on ecosystem processes, including carbon cycling, can be varied and result from competing driving factors controlling the response of soil respiration in different environments. For instance, soil respiration in a grassland was 20 % greater than in a forest under similar conditions (Raich and Tufekcioglu 2000) due to higher input of carbon in grassland, but the effect size of experimental warming on soil respiration was greater in the forest than in the grassland (Rustad et al. 2001). Previous studies have shown that experimental warming changes not only soil temperature, but also soil moisture (Dabros et al. 2010; Xu et al. 2010). Soil moisture affects soil respiration and its temperature sensitivity by changing oxygen availability or by alleviating substrate diffusion limitation (Moyano et al. 2012; Schmidt et al. 2004; Suseela et al. 2012). The reduced effect of experimental warming on soil respiration in grassland compared to forest could have been caused by a soil moisture limitation in the grassland, occurring either naturally or caused by experimental warming. Although no significant difference has been found between responses to various warming techniques (Rustad et al. 2001), most studies in the Arctic region have used open-top chambers, while most studies in grasslands have used infra-red radiator and passive nighttime warming chambers (Aronson and McNulty 2009). This suggests that varying responses between ecosystems might have been caused by methodological differences (Klein et al. 2005). The way different ecosystems respond to warming and the soil respiration response to those environmental changes must be studied using a single experimental warming technique in the different ecosystems. In addition, temperate and boreal forest ecosystems, regions at higher latitude with low precipitation, and arid and semi-arid biomes are underrepresented in experimental warming and soil respiration studies (Aronson and McNulty
2009; Raich and Schlesinger 1992; Rustad 2008). Most experimental warming studies have been conducted in North America and Europe, in mid to high latitudes and moderate to high annual precipitation. Study of climate change impacts on soil respiration in Northern Mongolia is of particular interest for two reasons. Our study site is located at the southern fringe of Siberian continuous permafrost and consists of three ecosystems in close proximity: Siberian boreal forest, tussocky peat, and Central Asian semi-arid steppe. This allows us to compare the responses of soil respiration to alteration of microclimatic conditions in different ecosystems that experience the same climatic variation. In addition, the observed temperature increase in Northern Mongolia between 1963 and 2002 has been 1.8 _C (Nandintsetseg et al. 2007), greater than global average temperature increases (IPCC 2007). At the same time, the forested area in Mongolia has significantly decreased as evapotranspiration increasingly exceeds current precipitation (Dagvadorj et al. 2009). In the future, this region is expected to undergo greater change than global average changes in climate (Dagvadorj et al. 2009; Meehl et al. 2007), and the forested area is projected to be replaced by steppe (Dagvadorj et al. 2009). This projected change is likely to have a strong impact on soil respiration and the overall carbon balance of the region (Cahoon et al. 2012). The objective of the current study was to determine the response of soil respiration to experimental warming, and to determine if this response differs among ecosystems. We manipulated air temperature, soil temperature and soil moisture using passive, opentop chambers (OTCs) and periodically measured soil respiration over the course of three growing seasons in three different ecosystem settings (i.e., boreal forest, riparian area, and semi-arid steppe) in a field site in Northern Mongolia. We anticipated that soil respiration would increase in response to the chamber treatment in the forest, riparian and steppe ecosystem, but had no a priori expectations for the magnitudes of these increases due to large differences in potentially competing effects among the ecosystems.

Material and Methods

Study site
The study site is located in the Dalbay Valley, in the Lake Hovsgol International Long-Term Ecological Research (ILTER) site, in Northern Mongolia (51_01.4050N, 100_45.6000E; 1,670 m asl). The mean annual temperature of this region is -4.5 _C, with the coldest average temperature of-21 _C in January, and the warmest average temperature of 12 _C in July (Nandintsetseg et al. 2007). The mean annual rainfall ranges between 290 and 300 mm in lower altitudes (Namkhaijantsan 2006).
The experiment was performed in three ecosystems located in close proximity to each other within Dalbay Valley: (1) semi-arid steppe, located on the south-facing slope, which is free of permafrost, (2) shrub-dominated riparian zone, located in the valley bottom with underlying permafrost, and (3) larch forest, with underlying permafrost, on the north-facing slope (Fig. 1). Of two commonly occurring trees, Siberian larch (Larix sibirica) and Siberian pine (Pinus sibirica), Siberian larch is the dominant tree in the forest. Average tree height is 10 m, average DBH is 15 cm, and average stand density is *3,500 trees ha-1. Dominant understory species in the forest are sub-shrub (Vaccinum vitis-idaea), moss (e.g., Rhytidium rugosum), grass (Festuca lenensis) and forbs (e.g., Galium boreale, Chrysanthemum zawadskii, Peucedanum sp.). The riparian zone where our experimental blocks were located is characterized by tall shrubs (Salix sp.) up to a height of 1.8 m and clear patches dominated by forbs (e.g., Artemisia tanacetifolia, Silene repense, Myosotis sylvatica), grass (e.g., Leymus chinensis, Poa subfastigiata, Agrostis mongolica) and sedges (e.g., Carex melanocephala, Carex sp.).
The foot of the south-facing slope of the valleywhere our steppe experimental blockswere located is dominated by sedges (Carex pediformis), grasses (e.g., F. lenensis, Helictotrichon schellianum, Koeleria macrantha), forbs (e.g., Potentilla acaulis, Aster alpinus, Artemisia commutata) and sub-shrubs (Thymus gobicus). The dominant soil type differs among the three ecosystems. According to (Batkhishig 2006) the forest soils were Mountain Taiga-derno in the Mongolian soil classification system (equivalent to a Cryept in U.S. Soil Taxonomy), alluvial meadow boggy cryomorphic soil (Fluvent) in the riparian area, and noncalcareous dark Kastanozem (Aridic Boroll or Typic Ustoll) in the steppe. Soil properties were determined from samples collected from quantitative soil pits excavated in 2010 and 2011 according to the methods of Hamburg (1984). Pits were dug to bedrock or 100 cm (whichever was deepest) and samples were collected for each 10 cm depth increment, although samples from only the surface (0–10 cm) layer were used in the current study. Analyses using standard methods showed that soil texture was relatively consistent among ecosystems, but that soils differed in pH, effective cation exchange capacity, and C and N contents among ecosystems (Table 1). Since the study area became a national park, land-use has been minimized, though the steppe on the south-facing slope has been used as grazing pasture and some parts of the riparian area has been used for hay preparation. Four replicate transects across the three ecosystems were established in mid-June of 2009, yielding 12 blocks (Fig. 1).



The distances between blocks are approximately 1 km in the same environmental setting and approximately 300–700 m along the cross-section of Dalbay Valley. International Tundra Experiment (ITEX)-style open-top passive warming chamber (OTC), measuring 1.5 m from side to side at the bottom, and 1.0 m at the top at 40 cm above the ground surface was used to manipulate climate (Marion et al. 1997). In each block, OTCs and a non-warmed control area with the same footprint were installed. The OTCs were consistently installed in the same locations for three growing seasons beginning in June of the summers of 2009, 2010 and 2011, and retrieved at the end of August of each year. Forest blocks were located under a closed larch forest canopy. Average ground cover inside OTCs and control forest plots was *75 % according to 2009 and 2011 plant cover estimates, and the vegetation was typical short-statured understory vegetation that did not reach the top of the chambers or create a microcanopy. Riparian blocks had no shrubs inside or near chambers or control plots, but had dense ground cover (nearly 100 % coverage). Vegetation grew taller than OTCs in two riparian blocks generating a significant micro-canopy within the chambers, but not in the other two plots. The steppe blocks were characterized by vegetation with short stature (*10 cm above ground), with sparse coverage (*68 %) that did not generate any significant micro-canopy within the chambers.
Environmental monitoring
Air temperature, soil temperature and soil moisture were measured to record changes in environmental variables in response to the chamber treatment. Above-ground air temperature (15 cm) was continuously recorded in each treatment (OTC and control) using HOBO pendant dataloggers (±0.5 _C accuracy; Onset Computers Inc., Bourne, MA, USA) at intervals of 30 min. The air temperature dataloggers were placed inside of RS3 radiation shields (Onset Computer Corporation, Pocasset, MA), which were placed in the middle of plot. Soil temperature and moisture were measured and recorded using calibrated EC-TM sensors (±1 _C and 1–3 %VWC accuracy) and EM50 dataloggers (Decagon Devices Inc., Pullman, WA, USA) at intervals of 30 min in each treatment of blocks. The soil temperature and moisture sensors were placed horizontally at depths of 10 cm. To determine how experimental warming and subsequent changes in environmental variables affect soil respiration, surface CO2 efflux was measured using a portable infra-red gas analyzer (IRGA; EGM- 4, PP Systems Inc.) and soil respiration chamber (SRC-1, PP Systems Inc.) in consistently the same locations, which were kept free of green and standing dead plants throughout the study period. Litter material was returned to the measurement location immediately after analysis to avoid changes in surface temperature, moisture and decomposition regime. Soil respiration was measured three times per treatment per block, and the mean was used for statistical analyses. Each measurement lasted 3 min. It was possible to measure only three blocks per day without introducing diurnal variation in the soil respiration measurements. Therefore, one block (out of four) was chosen randomly from each ecosystem, and these three blocks were measured in a given day. In subsequent days, additional sets of three blocks were randomly chosen (one from each ecosystem, with previously sampled blocks left out of the selection), until all 12 blocks were sampled in a four-day span before restarting the random selection process. The order of measurement of these three blocks was randomized to avoid a measurement order bias. During each growing season, 13–15 measurements were taken in each block.
Data analysis
Environmental data analysis

Daily sinusoidal fluctuations in measured air and soil temperatures were removed using Fourier transformation and applying high frequency filters (MATLAB v5, MathWorks Inc, Natick, MA) to identify outliers in the environmental data set caused by instrumental errors. Data points that fell outside of three standard deviations from mean-normalized data were considered erroneous and excluded from analysis. The proportion of erroneous temperature measurements was typically 0.6 % for air temperature data and 1.0–1.3 % for soil temperature data. A small proportion (�.5 %) of soil moisture data were negative values and thus considered erroneous and excluded. Mean daily values for each environmental variable were calculated from non-transformed outlier-free data and used for further statistical analyses.
 
Statistical tests of chamber treatment on environmental variables
Chamber effects on environmental variables (air temperature, soil temperature and moisture) and CO2 efflux rates were evaluated using repeated-measures ANOVA with ecosystems, chamber treatment, and all their interactions as fixed factors, and blocks as a random factor nested within ecosystems. Significant inter-annual variability was detected, and therefore the effects of the chamber treatment and ecosystems were evaluated for each year separately. When analyzing CO2 efflux response to chamber treatment, the fourday span required for measuring CO2 efflux in all replicate blocks was considered the time unit for the repeated measures analyses. Mean daily values were time units for other analyses. When differences among ecosystems and interactions were statistically significant, these differences were tested using Tukey’s HSD test. All ANOVA analyses were carried out with JMP v8 (SAS Institute, Cary, NC). The design and results of these statistical analyses are reported in the Supplementary Materials, while P values for comparisons are reported in the main text. The relationship between soil temperature at 10 cm depth and soil respiration was modeled by fitting an exponential function to the OTC and control data of each ecosystem, with data from all years combined:

 

Rij = b0eb1Tij
Where Rij is the soil respiration rate (lmol CO2 m-2 s-1) in either chamber or control plot (i) of one of the ecosystems (j), Tij is the soil temperature (0C) at 10 cm depth recorded at the same time as the respiration measurement, b0 is the modeled intercept of soil respiration, and b1 is themodeled temperature sensitivity coefficient. The regression parameters were then used to calculate apparent Q10 values of each data set using the following equation:
Q10 = e10b1
In addition, the relationship between soil moisture at 10 cm and soil respiration was modeled by linear regression, with data from all years combined. Regression curve fitting and corresponding parameters and goodness-of-fit tests were carried out using SigmaPlot v12 (Systat Software Inc. San Jose, CA).
 
Statistical analysis of competing effects on soil respiration
An information-theoretic approach was used to assess the relative importance of environmental variables, ecosystem type and year on soil respiration. The same data set used for ANOVA tests of chamber effects were re-analyzed using Akaike’s information criterion, which is corrected for the number of parameters and sample size (AICc), and evidence ratios to determine the best supported model (Burnham and Anderson 2002). Evidence ratios provide an assessment of two competing models, where larger ratios indicate more favored models. Another measure used to compute the set of the best models is Akaike weights (i.e., model probabilities) of the ranked models which were summed until a cutoff point of 95 % is reached. Because of interactions involving years and ecosystems, competing models were examined in two ways: by ecosystem with all annual data combined, as well as by year with all ecosystems combined. We used main effects models with soil temperature and moisture as continuous variables, and chamber effect, ecosystem and years as categorical variables. To examine the relative importance of each environmental variable, standardized partial regression coefficients (b0) were computed for the best supported models. Model selection and estimation of regression coefficients were carried out using the R statistical package (R Development Core Team 2011) with the AICcmodavg package (Mazerolle 2012).

Result

Environmental variables
Chamber effect on air temperature
Mean seasonal ambient air temperatures differed among the forest, riparian and steppe areas (P<0.001;Table S1) in the order steppe[riparian[forest, and among the 3 years (P<0.001) in the order 2009>2010>2011 (Fig. 2). OTCs significantly increased air temperature in 2010 and 2011 (P<0.01; Table S1) but not in 2009 (P = 0.25; Table S1) across the three ecosystems (Fig. 2). However, the magnitude of the increase in air temperature by OTCs differed among ecosystems (P<0.05 for OTC 9 Ecosystem; Table S1). The air temperature increase by OTCs was greatest in the steppe (1.0–2.1 _C in 2010 and 2011; P<0.05, Table S1), intermediate in the riparian zone (0.5–0.6 _C; P = 0.1– 0.2; Table S1), and smallest in the forest (0.2–0.4 _C; P = 0.4–0.6; Table S1). The magnitude of warming by OTCs significantly decreased in the riparian area aftermid of June (P<0.001; Table S1), but this trend was not consistently observed in the forest and steppe (see Supplementary Materials).
Chamber effect on soil temperature
Similar to air temperature, mean daily soil temperatures at 10 cm depth were lowest in the forest, followed by the riparian zone, and were greatest in the steppe (P<0.001; Table S2; Fig. 3).

The chamber effect was statistically significant only in 2010 (P<0.05; Table S2), and no interaction effect between ecosystem and chamber treatment was observed (P = 0.1–0.5; Table S2; Fig. 3). In the riparian zone, soil temperature differences due to the chamber effect were highly variable (positive in some blocks and negative in others) and therefore no overall difference was observed (P = 0.6–0.9; Table S2). In contrast, the seasonal mean soil temperatures were greater in OTCs than in controls in all blocks of the forest (by 0.6–1.4 0C) and the steppe (by 1.0–1.7 0C). However, the differences were statistically significant only in the steppe (P<0.05 in the steppe, and P = 0.1–0.2 in the forest; Table S2).


 
Chamber effect on soil moisture
Soil moisture differed significantly among ecosystems (P<0.05;Table S3), where itwas greatest in the riparian area, followed by the forest, and was least in the steppe (Fig. 4). Due to high spatial variability among replicate blocks (65 % of total variance) and of chamber treatments 9 blocks (26–28 %), chambers had no statistically significant effect on soil moisture (P = 0.06–0.9;
Table S3). However, soilmoisture was less in OTCs than in control of the steppe (an absolute decrease of 3.0–6.2 % VWC) and riparian area (1.6–2.3 % VWC in 2009 and 2010, and 11.1 %VWCin 2011), but greater in OTCs in the forest (by 3.9–10.6 ± 0.6 % VWC in 2009 and 2010, with no difference in 2011; Fig. 4).
 
Soil respiration
Soil respiration variation across ecosystems and years
Soil respiration rates in the control plots varied significantly across ecosystems (P<0.05; Table S4) and years (P<0.05) (Figs. 5, 6). The largest soil respiration rates were observed in the riparian area (5.9 ± 0.2 lmol CO2 m-2 s-1), followed by the forest (4.9 ± 0.1 lmol CO2 m-2 s-1), and was least in the steppe (3.2 ± 0.1 lmol CO2 m-2 s-1). However, when normalized to soilCcontent in the 0–10 cm layer (Table 1), mean respiration rates in the riparian and steppe areas were similar, while respiration in the forest was slightly higher (0.16 mg CO2–C g-1 soil C h-1 in riparian, 0.17 mg CO2–C g-1 soil C h-1 in steppe and 0.20 mg CO2–C g-1 soil C h-1 in forest). Soil respiration rates were greatest across all ecosystems in 2009 than in other years (19–20 %greater than the average of all years), were least in 2010 (11–23 % less than average), and varied considerably in 2011 (10 % less than average in the forest, but 15 and 2 % greater than average in the riparian area and the steppe, respectively). The greater variability can be attributed to much fewer measurements made in 2011 due to equipment problems.



 
Chamber effect on soil respiration
The OTC had no significant effect on soil respiration in any ecosystem (P = 0.7–0.9; Table S4). However, soil respiration rates in the forest were greater by 1.1 lmol CO2 m-2 s-1 in OTCs than in controls in 2009, but the difference was not statistically significant (P = 0.15; Table S4). In the riparian and steppe areas, soil respiration rates in OTCs and controls were similar in all years (P = 0.2–0.9; Table S4; Fig. 5). The lack of differences between OTCs and controls in these ecosystems is attributable in part to high variability among the plots, which represented 16–41 % of the variance directly and 20–70 % of the variance through interaction effects.
 
Temperature sensitivity of soil respiration
The temperature sensitivity of soil respiration varied widely among treatments and ecosystems, with estimated Q10 values ranging from 1.3 to 5.8 (Table 2; Fig. 7). The 95 % confidence intervals of temperature sensitivity coefficients (b1) showed that the coefficients of the forest plots were consistently greater than those of the riparian and steppe plots, with the exception of the control-steppe plots. In the forest, control plots had a significantly greater temperature sensitivity coefficient than OTCs, such that the lower 95 % confidence interval of the controls did not overlap with upper confidence limit of the OTCs. These coefficients yielded greater apparent Q10 values for control plots than OTCs (Table 2). Conversely, 95 % confidence limits of temperature sensitivity coefficients of control and OTC treatments overlapped with each other in both the riparian and steppe blocks (Table 2), suggesting that the chamber treatment did not alter temperature sensitivity in these ecosystems.


 
Best supported models for relationships between soil respiration and environmental variables
 
The best supported models from data separated by ecosystem with annual data combined included soil temperature, soil moisture, chamber treatment and year for the forest and steppe ecosystems (Table 3). When effects of soil temperature and moisture on soil respiration were compared in each ecosystem, soil temperature and moisture appear to have similar and equivalent effects on soil respiration in the steppe (b0 = 0.38 for temperature and b0 = 0.41 for soil moisture). In contrast, the soil temperature effect was greater than the soil moisture effect in the forest, which is demonstrated by relatively stronger positive correlation with soil temperature (b0 = 0.46) versus relatively weak negative correlation with soil moisture (b0 = -0.17). The best supported model for the riparian ecosystem included soil temperature, moisture and year (Table 3). In the riparian ecosystem, soil temperature and moisture also appear to have similar effects on soil respiration (b0 = 0.55 for temperature and b0 = 0.45 for soil moisture). When data were modeled by year with ecosystems combined, the best supported model for 2009 included soil temperature and ecosystems only, and the best supported models for 2010 and 2011 included soil temperature, soil moisture and ecosystems (Table 4). In 2009, which was the wettest and coolest year of the study, the top four models (representing the 95 % confidence set on the best supported models) all contained soil temperature. Conversely, in 2011, which was the driest and warmest year of the study, the 95 % confidence set on the best supported models also contained four models, and all models included soil moisture (Table 4). The partial regression coefficients for soil temperature were greater than for soil moisture in the colder and wetter years (2009: b0 = 0.8 for temperature vs b0 = 0.1 for moisture; 2010: b0 = 1.1 for temperature vs b0 = 0.4 for moisture). Conversely, the partial regression coefficients for soil temperature and soil moisture similar in the warmer and drier year of 2011 (b0 = 0.4 for both temperature and moisture).

Discussion

Chamber effects on environmental variables
To the best of our knowledge, there are three published studies using OTCs in forested systems (De Frenne et al. 2010; Sjo¨gersten and Wookey 2009; Xu et al. 2010). Compared to these studies, OTCs in our forest plots generated smaller air temperature increases and larger soil temperature increases. Air and soil warming by OTCs in the steppe was also greater than previous studies performed in similar open systems with short stature vegetation (Carlyle et al. 2011; Kudernatsch et al. 2008). Soil moisture decreases in OTCs in the steppe and riparian blocks can be attributed mainly to reductions in incident rainfall and were consistent with other studies (Carlyle et al. 2011; Kudernatsch et al. 2008; Xia et al. 2009). In contrast, soil moisture was greater in OTCs in the forest compared to control plots. De Frenne et al. (2010) also reported soil moisture increases in OTCs, though the amount of soil moisture increase was smaller than in our study. Overall, the observed differences in the responses of environmental variables to a single experimental warming treatment applied simultaneously to forest, riparian and steppe ecosystems clearly demonstrate the importance of ecosystem setting and antecedent conditions (Shaver et al. 2000).






Such conditions include the openness of the canopy surrounding the blocks, the stature of the vegetation inside the chambers, and the slope and aspect of the ground surface. All of these differed among the ecosystems and likely affected the energy and water balances of the systems through differences in incident solar radiation, evapotranspiration or convective airflow.
 
Competing effects on soil respiration
Our results showed that seasonal mean soil respiration rates differed significantly among ecosystems, but that these differences are attenuated when respiration rates were normalized to soil carbon content. The normalized soil respiration rate was slightly higher in forest, suggesting more favorable microclimate conditions (Kang et al. 2003; Rustad et al. 2001; Sjo¨gersten and Wookey 2002). Although all ecosystems shared the same long-termmean annual temperature and precipitation based on regional weather data, the closed forest canopy coupled with underlying permafrost could have created more favorable microclimate and reduced moisture limitation stress in the forest and thus had a strong indirect positive effect on soil respiration during growing season (Cahoon et al. 2012; Conant et al. 1998; Matı´as et al. 2012). Regression analyses demonstrated that soil respiration was strongly related to both soil temperature and moisture, and competing models analyses demonstrated that the most important factor governing soil respiration varied among ecosystems. Soil temperature was a stronger predictor of soil respiration in the forest, consistent with previous studies conducted in woody ecosystems (Bergner et al. 2004; Cahoon et al. 2012; German et al. 2012; Pan et al. 2008;Xuet al. 2010), while soil moisture was the more important determinant of soil respiration in the steppe (Lellei-Kovacs et al. 2008; Liu et al. 2009; Matı´as et al. 2012). Application of the chamber treatments across ecosystems resulted in confounded alterations of environmental variables, butwere able to demonstrate that soil respiration in the forest may increase in response to climate change due to greater temperature sensitivity, but soil respiration in the steppe may not respond to the same degree because of less temperature sensitivity, which was likely caused by soil moisture limitation. These divergent responses to the chamber treatment demonstrate that vegetation types need to be taken into account to improve regional model for carbon dynamics responses to climatic change. Ecosystem boundaries may change due to direct and indirect effects of climate change, further altering the potential carbon balance of a given area. Normalized Difference Vegetation Index (NDVI) data showed that desert and steppe areas have increased, and forest area has decreased in Mongolia (Dagvadorj et al. 2009). Indeed, tree-ring width and regeneration of Siberian larch have decreased since 1940 as aridity and mean annual air temperature have increased in Mongolia (Dulamsuren et al. 2010). According to regional models, these current trends of ecosystem shifts may continue in the future (Batima et al. 2005; Dulamsuren et al. 2010) due to projected aridity intensification (Batima et al. 2005; Sato et al. 2007). Lu et al. (2009) modeled C dynamics in Mongolia and concluded that this region was a sink of 31 Tg C year-1 in the 1990s. However, this sink is highly vulnerable because of the significant decreases in aboveground C stocks and potential decreases in belowground C stocks typically anticipated when steppe replaces forest. Further carbon losses may be anticipated due to enhanced respiration in response to warming, but the magnitude of the increase is difficult to predict at the landscape or regional scale. Our results suggest that changes in precipitation may have stronger effects than warming on soil respiration in the steppe by changing soil moisture availability and indirectly by affecting primary productivity (Knapp et al. 2002), substrate supply and drought stress (Davidson et al. 2006). While outside the scope of the current study, important questions remain about how primary productivity would respond to temperature and moisture change in these systems, and whether an increase in carbon uptake can offset carbon loss due to enhanced ecosystem respiration. Studies conducted in similar arid ecosystems reported that the ecosystem switched from sink to source when water stress was observed (Fu et al. 2006; Niu et al. 2008). Our results thus highlight the importance of identifying the relevant environmental factors that govern soil respiration in different ecosystems, as well as the sensitivity of soil respiration to changes in those factors for predicting potential landscape-scale changes in carbon cycling in response to anticipated climate change.

Acknowledgement

We thank J. Mortensen, D. Brickley, S. Undrakhbold, J. Batbaatar, N. Sandag, research camp staff, and the contributing American and Mongolian undergraduates for their assistance throughout the project. The study was conducted within the framework of PIRE-Mongolia project, supported by National Science Foundation Grant OISE 0729786.

Reference

  1. Aronson EL, McNulty SG (2009) Appropriate experimental ecosystem warming methods by ecosystem, objective, and practicality. Agric For Meteorol 149(11):1791–1799
  2. Batima P, Natsagdorj L, Gombluudev P, Erdenetsetseg B (2005) Observed climate change in Mongolia. AIACC working paper No. 12, p 25
  3. Batkhishig O (2006) Soils of the Lake Hovsgol area and its watershed. In: Goulden CE, Sitnikova T, Gelhaus J, Boldgiv B (eds) The geology, biodiversity and ecology of Lake Hovsgol (Mongolia). Biology of inland waters. Backhuys Publisher, Leiden, pp 1–20
  4. Bergner B, Johnstone J, Treseder KK (2004) Experimental warming and burn severity alter soil CO2 flux and soil functional groups in a recently burned boreal forest. Glob Change Biol 10(12):1996–2004
  5. Biasi C, Meyer H, Rusalimova O, Ha¨mmerle R, Kaiser C, Baranyi C, Daims H, Lashchinsky N, Barsukov P, Richter A (2008) Initial effects of experimental warming on carbon exchange rates, plant growth and microbial dynamics of a lichen-rich dwarf shrub tundra in Siberia. Plant Soil, 307(1–2):191–205
  6. Bond-Lamberty B, Thomson A (2010) Temperature-associated increases in the global soil respiration record. Nature 464(7288):U132–U579
  7. Burnham KP, Anderson DR (2002) Model selection and multimodel inference. Springer, New York
  8. Cahoon SMP, Sullivan PF, Shaver GR, Welker JM, Post E (2012) Interactions among shrub cover and the soil microclimate may determine future Arctic carbon budgets. Ecol Lett 15(12):1415–1422
  9. Carlyle C, Fraser L, Turkington R (2011) Tracking soil temperature and moisture in a multi-factor climate experiment in temperate grassland: do climate manipulation methods produce their intended effects? Ecosystems 14(3):489–502
  10. Conant RT, Klopatek JM, Malin RC, Klopatek CC (1998) Carbon pools and fluxes along an environmental gradient in Northern Arizona. Biogeochemistry 43(1):43–61
  11. Curtis PS, Vogel CS, Gough CM, Schmid HP, Su HB, Bovard BD (2005) Respiratory carbon losses and the carbon-use efficiency of a northern hardwood forest, 1999–2003. New Phytol 167(2):437–456
  12. Dabros A, Fyles JW, Strachan IB (2010) Effects of open-top chambers on physical properties of air and soil at postdisturbance sites in northwestern Quebec. Plant Soil 333(1):203–218
  13. Dagvadorj D, Natsagdorj L, Dorjpurev J, Namkhainyam B (2009) Mongolia assessment report on climate change 2009. Hiimori Printing, Ulaanbaatar, p 228
  14. Davidson EA, Janssens IA, Lou Y (2006) On the variability of respiration in terrestrial ecosystems: moving beyond Q10. Glob Chang Biol 12(2):154–164
  15. De Frenne P, De Schrijver A, Graae BJ, Gruwez R, Tack W, Vandelook F, Hermy M, Verheyen K (2010) The use of open-top chambers in forests for evaluating warming effects on herbaceous understorey plants. Ecol Res 25(1):163–171
  16. Dulamsuren C, Hauck M, Leuschner C (2010) Recent drought stress leads to growth reductions in Larix sibirica in the western Khentey, Mongolia. Glob Change Biol 16(11): 3024–3035
  17. Fu Y-L, Yu G-R, Sun X-M, Li Y-N, Wen X-F, Zhang L-M, Li Z-Q, Zhao L, Hao Y-B (2006) Depression of net ecosystem CO2 exchange in semi-arid Leymus chinensis steppe and alpine shrub. Agric For Meteorol 137(3–4):234–244
  18. German DP, Marcelo KRB, Stone MM, Allison SD (2012) The Michaelis–Menten kinetics of soil extracellular enzymes in response to temperature: a cross-latitudinal study. Glob Change Biol 18(4):1468–1479
  19. Hamburg SP (1984) Forest soils and treatment impacts. In: Stone EL (ed) North American Forest Soils Conference, Society of American Foresters Publication, Department of Forestry, Wildlife and Fisheries, University of Tennessee, Knoxville
  20. Heimann M, Reichstein M (2008) Terrestrial ecosystem carbon dynamics and climate feedbacks. Nature 451(7176): 289–292
  21. IPCC (2007) Climate change 2007: synthesis report. Contribution of Working Groups I, II and III to the fourth assessment report of the intergovernmental panel on climate change. Geneva, Switzerland
  22. Kang S, Doh S, Lee D, Jin VL, Kimball JS (2003) Topographic and climatic controls on soil respiration in six temperate mixed-hardwood forest slopes, Korea. Glob Change Biol 9(10):1427–1437
  23. Klein JA, Harte J, Zhao XQ (2005) Dynamic and complex microclimate responses to warming and grazing manipulations. Glob Change Biol 11(9):1440–1451
  24. Knapp AK, Fay PA, Blair JM, Collins SL, Smith MD, Carlisle JD, Harper CW, Danner BT, Lett MS, McCarron JK (2002) Rainfall variability, carbon cycling, and plant species diversity in a mesic grassland. Science 298(5601): 2202–2205
  25. Kudernatsch T, Fischer A, Bernhardt-Ro¨mermann M, Abs C (2008) Short-term effects of temperature enhancement on growth and reproduction of alpine grassland species. Basic Appl Ecol 9(3):263–274
  26. Lellei-Kovacs E, Kovacs-Lang E, Kalapos T, Botta-Dukat Z, Barabas S, Beier C (2008) Experimental warming does not enhance soil respiration in a semiarid temperate forest steppe ecosystem. Community Ecol 9(1):29–37
  27. Liu WX, Zhang Z, Wan SQ (2009) Predominant role of water in regulating soil and microbial respiration and their responses to climate change in a semiarid grassland. Glob Change Biol 15(1):184–195
  28. Lu Y, Zhuang Q, Zhou G, Sirin A, Melillo J, Kicklighter D (2009) Possible decline of the carbon sink in the Mongolian Plateau during the 21st century. Environ Res Lett 4(4): 045023
  29. Marion GM, Henry GHR, Freckman DW, Johnstone J, Jones G, Jones MH, Le´vesque E, Molau U, Mølgaard P, Parsons AN, Svoboda J, Virginia RA (1997) Open-top designs for manipulating field temperature in high-latitude ecosystems. Glob Change Biol 3(S1):20–32
  30. Matı´as L, Castro J, Zamora R (2012) Effect of simulated climate change on soil respiration in a mediterranean-type ecosystem: rainfall and habitat type are more important than temperature or the soil carbon pool. Ecosystems 15(2): 299–310
  31. Mazerolle MJ (2012) AICcmodavg. R package version 1.24. In: R Foundation for Statistical Computing. Vienna, Austria
  32. Meehl GA, Stocker TF, Collins WD, Friedlingstein P, Gaye AT, Gregory JM, Kitoh A, Knutti R, Murphy JM, Noda A, Raper SCB, Watterson IG, Weaver AJ, Zhao Z-C (2007). Global climate projections. In: Solomon
  33. S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL (eds) Climate change 2007: The physical science basis. Contribution of Working Group I to the fourth assessment report of the intergovernmental panel on climate change. Cambridge, UK and New York, NY, USA
  34. Moyano FE, Vasilyeva N, Bouckaert L, Cook F, Craine J, Curiel Yuste J, Don A, Epron D, Formanek P, Franzluebbers A, Ilstedt U, Ka¨tterer T, Orchard V, Reichstein M, Rey A, Ruamps L, Subke JA, Thomsen IK, Chenu C (2012) The moisture response of soil heterotrophic respiration: interaction with soil properties. Biogeosciences 9(3):1173– 1182
  35. Namkhaijantsan G (2006) Climate and climate change of the Hovsgol region. In: Goulden EC, Sitnikova T, Gelhaus J, Boldgiv B (eds) The geology, biodiversity and ecology of Lake Hovsgol (Mongolia). Backhuys Publisher, Leiden, pp 63–76
  36. Nandintsetseg B, Greene JS, Goulden CE (2007) Trends in extreme daily precipitation and temperature near lake Ho¨vsgo¨l, Mongolia. Int J Climatol 27(3):341–347
  37. Niu S, Wu M, Han Y, Xia J, Li L, Wan S (2008) Water-mediated responses of ecosystem carbon fluxes to climatic change in a temperate steppe. New Phytol 177(1):209–219
  38. Pan XL, Lin B, Liu Q (2008) Effects of elevated temperature on soil organic carbon and soil respiration under subalpine coniferous forest in western Sichuan Province, China. Chin J Appl Ecol 19(8):1637–1643
  39. R Development Core Team (2011) R: a language and environment for statistical computing, reference index version 2.13. In: R Foundation for Statistical Computing. Vienna, Austria
  40. Raich JW, Schlesinger WH (1992) The global carbon-dioxide flux in soil respiration and its relationship to vegetation and climate. Tellus B Chem Phys Meteorol 44(2):81–99
  41. Raich JW, TufekciogluA(2000) Vegetation and soil respiration: correlations and controls. Biogeochemistry 48(1):71–90
  42. Rustad LE (2008) The response of terrestrial ecosystems to global climate change: towards an integrated approach. Sci Total Environ 404(2–3):222–235
  43. Rustad LE, Campbell JL, Marion GM, Norby RJ, Mitchell MJ, Hartley AE, Cornelissen JHC, Gurevitch J, Gcte N(2001)A meta-analysis of the response of soil respiration, net nitrogen mineralization, and aboveground plant growth to experimental ecosystem warming. Oecologia 126(4):543–562
  44. Sato T, Kimura F, Kitoh A (2007) Projection of global warming onto regional precipitation over Mongolia using a regional climate model. J Hydrol 333(1):144–154
  45. Schlesinger WH (1977) Carbon balance in terrestrial detritus. Annu Rev Ecol Syst 8:51–81
  46. Schmidt IK, Tietema A, Williams D, Gundersen P, Beier C, Emmett BA, Estiarte M(2004) Soil solution chemistry and element fluxes in three European heathlands and their responses to warming and drought. Ecosystems 7(6):638– 649
  47. Shaver GR, Canadell J, Chapin FS III, Gurevitch J, Harte J, Henry G, Ineson P, Jonasson S, Melillo J, Pitelka L, Rustad L (2000) Global warming and terrestrial ecosystems: a conceptual framework for analysis. BioScience 50(10): 871–882
  48. Sjo¨gersten S, Wookey PA (2002) Climatic and resource quality controls on soil respiration across a forest-tundra ecotone in Swedish Lapland. Soil Biol Biochem 34(11):1633–1646
  49. Sjo¨gersten S, Wookey PA (2009) The impact of climate change on ecosystem carbon dynamics at the Scandinavian mountain birch forest-tundra heath ecotone. Ambio 38(1): 2–10
  50. Suseela V, Conant RT, Wallenstein MD, Dukes JS (2012) Effects of soil moisture on the temperature sensitivity of heterotrophic respiration vary seasonally in an old-field climate change experiment. Glob Change Biol 18(1):336– 348
  51. Wan S, Hui D, Wallace L, Luo Y (2005) Direct and indirect effects of experimental warming on ecosystem carbon processes in a tallgrass prairie. Global Biogeochem Cycles 19(2):1–13
  52. Wan S, Norby RJ, Ledford J, Weltzin JF (2007) Responses of soil respiration to elevated CO2, air warming, and changing soil water availability in a model old-field grassland. Glob Change Biol 13(11):2411–2424 Xia JY, Niu SL,
  53. Wan SQ (2009) Response of ecosystem carbon exchange to warming and nitrogen addition during two hydrologically contrasting growing seasons in a temperate steppe. Glob Change Biol 15(6):1544–1556
  54. Xu Z, Wan C, Xiong P, Tang Z, Hu R, Cao G, Liu Q (2010) Initial responses of soil CO2 efflux and C, N pools to experimental warming in two contrasting forest ecosystems, Eastern Tibetan Plateau, China. Plant Soil 336(1–2): 183–195
  55. Zhang PC, Tang YH, Hirota M, Yamamoto A, Mariko S (2009) Use of a regression method to partition sources of ecosystem respiration in an alpine meadow. Soil Biol Biochem 41(4):663–670