Altitude pattern of carbon stocks in desert grasslands of an arid land region

Rong Yang , JunQia Kong, ZeYu Du, YongZhong Su

Linze Inland River Basin Research Station, Key Laboratory of Ecohydrology of Inland River Basin, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Gansu 730000, China

ABSTRACT For estimating the altitude-distribution pattern of carbon stocks in desert grasslands and analyzing the possible mechanism for this distribution, a detailed study was performed through a series of field vegetation surveys and soil samplings from 90 vegetation plots and 45 soil profiles at 9 sites of the Hexi Corridor region, Northwestern China. Aboveground, belowground, and litter-fall biomass-carbon stocks ranged from 43 to 109, 23 to 64, and 5 to 20 g/m2, with mean values of 80.82,44.91, and 12.15 g/m2, respectively. Soil-carbon stocks varied between 2.88 and 3.98 kg/m2, with a mean value of 3.43 kg/m2 in the 0–100-cm soil layer. Both biomass- and soil-carbon stocks had an increasing tendency corresponding to the altitudinal gradient. A significantly negative correlation was found between soil-carbon stock and mean annual temperature, with further better correlations between soil- and biomass-carbon stocks, and mean annual precipitation. Furthermore, soil carbon was found to be positively correlated with soil-silt and -clay content, and negatively correlated with soil bulk density and the volume percent of gravel. It can be concluded that variations in soil texture and climate condition were the key factors influencing the altitudinal pattern of carbon stocks in this desert-grassland ecosystem. Thus, by using the linear-regression functions between altitude and carbon stocks, approximately 4.18 Tg carbon were predicted from the 1,260 km2 of desert grasslands in the study area.

Keywords: altitudinal gradient; soil organic carbon; biomass carbon; soil bulk density; desert grasslands

1 Introduction

Grasslands are an important component of terrestrial ecosystems and play a crucial role in the global carbon (C) cycle (Scurlock and Hall, 1998). The grasslands in China are the third largest in the world and cover 4.0 × 108hm2(Fanget al., 2010). C-stock dynamics in the grassland ecosystems of China have received considerable research attention in the past ten years (Maet al., 2010; Yanget al., 2010).However, because of limited field observations and large spatial heterogeneity, the C stocks and distribution patterns in the grassland ecosystems of China remain largely uncertain (Fanget al., 2010). Therefore,for accurate assessment of the C stocks in the grassland ecosystems of China, it is necessary to obtain further basic soil and vegetation data through extensive field investigation.

Environmental gradients provide an excellent opportunity for both understanding the mechanisms of abiotic control on ecology processes and studying the potential impacts upon these processes (Kochet al.,1995; Saizet al., 2012). Altitudinal gradients are among the most powerful "natural experiments" for testing ecological and evolutionary responses of biota to environmental changes (K?rner, 2007). Altitude has profound effects on climatic and edaphic conditions (M?nnelet al., 2007), which have been identified as the principal factors influencing the C-stock distribution within ecosystems (Saizet al., 2012).Despite several studies addressing the C-stock-distribution pattern along altitudinal gradients in various ecosystems (Zhuanget al., 2007; Xuet al., 2010), this topic is yet not to be adequately addressed in the desert-grassland ecosystems of arid regions.

Desert grasslands are one of the major ecological landscapes in the Hexi Corridor region, and their conservation is important to provide ecosystem services,including C sequestration (Lüet al., 2014). Although a previous study showed a strong relationship between soil-organic-carbon (SOC) content and altitude in the grassland ecosystem of this area (Yanget al., 2014), it is difficult to study the mechanism responsible for altitudinal distribution patterns of ecosystem-C stocks across large spatial scales—because of the geography and vegetation heterogeneity and the differing frequency of disturbance events. However, it remains possible to extract valuable information about variations in C stocks along the altitudinal gradient of desert-grassland ecosystems across smaller spatial scales. Within the Hexi Corridor region, many desert-grassland ecosystems are distributed across the alluvial–diluvial fans of the Qilian Mountains. Therefore,a rich pattern of altitude distributions exists amongst these grasslands. Within a relatively small area in this region, we identified regular altitudinal gradients of uniform soil and vegetation conditions, which presented an ideal platform for the study of relationships between altitude and ecosystem-C stocks. Hence, we hypothesized that biomass-, soil-, and ecosystem-C stocks in desert grasslands would present regular distribution patterns along an altitudinal transect of the Hexi Corridor region. We expected to find further evidence of the contributing factors for the altitudinal distribution pattern of C stocks, such as soil and climate heterogeneity. We propose that the findings from this study could be used to estimate the ecosystem-C stocks of desert grasslands at various altitudes.

2 Materials and methods

2.1 Study sites

The study area is located in the middle of the Hexi Corridor region, Northwest China, and contains an extensive area of desert grasslands, with most of them distributed along an altitude transect on the northern piedmont of the Qilian Mountains (Figure 1). The study area has a typical desert climate, with windy and dry winters and springs, and warm and comparatively rain-rich summers followed by short and cool autumns (Yanget al., 2014). The desert grasslands studied shared a similar soil type, which was identified as a Calcic-Orthic Aridosol by the Chinese soil-taxonomy classification system (Cooperative Research Group on Chinese Soil Taxonomy, 2001) and is equivalent to an Aridosol under the United States Department of Agriculture soil-taxonomy classification (Soil Survey Staff, 2010). The predominant plant species in the desert grasslands of the study area are the subshrubsSalsola passerinaBunge andSympegma regeliiBunge.

Figure 1 Desert-grasslands distribution and sampling sites in the study area

2.2 Sampling design and chemical analysis

A detailed study was performed from July to October 2011 through a series of field vegetation surveys and soil samplings. Nine desert-grassland sites with consistent soil type and vegetation composition were selected, and some details about these sites are listed in Table 1. Ten quadrats of 1m × 1m were set up along a 100-m sampling transect at each of the sites, and numbered 1 through 10. In each of the quadrats, the aboveground biomass (AGB) and litter-fall biomass (LFB) were estimated by the harvest method. In quadrats 1, 3, 5, 7, and 9 (five quadrats),belowground biomass (BGB) was estimated by ex-cavating all plant roots from the 0–100-cm soil layer.AGB, BGB, and LFB samples were dried at 65 °C for 48 h in a hot-air oven and weighed for biomass measurement. These plant samples were then ground to＜0.5 mm for C-content analysis. Soil samples were taken from quadrats 1, 3, 5, 7, and 9 at depths of 0–5,5–10, 10–20, 20–30, 30–50, 50–70, and 70–100 cm.In quadrat 5, 1-m-deep soil was excavated; and five cutting-ring samples corresponding to the following layers were obtained to determine the soil bulk density (BD): 0–5, 5–10, 10–20, 20–30, 30–50, 50–70,and 70–100 cm. Soil samples were air-dried at room temperature and passed through a 2-mm sieve; and plant residues and visible organisms were removed.Gravel (＞2 mm) was weighed, and the volumetric gravel percent (VGP) of the soil obtained. The sieved soil samples were then separated into two parts. One aliquot was used to determine the soil-particle-size distribution. Subsamples were further ground and passed through a 0.25-mm sieve; the resultant sample was used for the SOC-content analysis.

Table 1 Aboveground biomass percentage of species at nine sampling sites with ten replicates

Soil-particle-size distribution was determined by the pipette method in a sedimentation cylinder by using sodium hexametaphosphate as the dispersing agent (Gee and Bauder, 1986). SOC content was determined by the dichromate oxidation method of Walkley–Black (Nelson and Sommers, 1973). C content in AGB, BGB, and LFB samples was measured using an elemental analysis instrument (Elementar vario MICRO cube, Germany).

2.3 Meteorological data, digital elevation data, and C-stock calculation

Datasets of mean annual temperature (MAT) and mean annual precipitation (MAP) were derived from the climatic data of the Hexi Corridor during 2011–2012 (http://cdc.cma.gov.cn). These data were spatially interpolated from the records of 25 meteorological stations around the study area. Data corresponding to the nine sampling sites were obtained through the interpolation map. The data set of the desert-grassland distributions in the study area in 2000 was provided by Environmental and Ecological Science Data Center for West China, National Natural Science Foundation of China (http://westdc.westgis.ac.cn).The digital elevation model data were downloaded from the International Scientific Data Service Platform (http://datamirror.csdb.cn/index.jsp) (Figure 1).

The biomass-carbon (BC) stocks were calculated by using the data from analyses of AGB, BGB, and LFB and the C content. The expression of SOC-stock calculation (kg/m2) is defined as follows:

where Hiis the thickness of the ith soil layer (cm);iis the ith soil layer; SOCiis the SOC content (g/kg) at layeri; BDiis the soil bulk density (g/cm3); and VPGithe volume percent of gravel ＞2 mm at theith layer (%).Total carbon (TC) stocks were given as the sum of BC and SOC stocks in the top 1 m.

2.4 Data analysis

Mean biomass, biomass-C content, and biomass-C stock at each sampling site were calculated from ten AGB and LFB replicates and five BGB replicates.One-way ANOVA was used to compare the difference significance of the variables. Mean biomass, biomass-C content, and biomass-C stock at nine sampling sites were the variables, while biomass compositions were the factors. When one-way analysis of variance (ANOVA) revealed a significant difference between the means, Tukey's honestly significant difference (HSD) test was performed for multiple comparisons between the means. Regression analyses were performed to evaluate the relationships between the soil-, biomass-, and ecosystem-C stocks and the altitude, and also between the C stocks and climate factors. A function, including the independent variable of altitude and the dependent variable of C stocks, was built using linear-regression analyses to predict the C stocks in desert grasslands at the ecosystem scale. With the data obtained from 315 soil samples in 45 quadrats of 9 sampling sites, partial correlation analysis was used to examine the relationship among SOC content, BD, VGP, and soil-particle size, which were variables, with soil depth as the control factor. These statistical analyses were conducted using SPSS, ver.19.0 (SPSS Inc., Chicago, IL, USA).The areas of desert grasslands distributed at a certain altitude were calculated using an area statistical function of ArcGIS software (ESRI, USA).

3 Results

3.1 Biomass-carbon stocks

AGB exhibited moderate variability among the different sample sites, ranging from 125 to 297 g/cm2.BGB and LFB exhibited large variability, ranging from 54.0 to 217 g/cm2for BGB and from 19 to 46 g/cm2for LFB (Table 2). C content of ABG, BGB, and LFB exhibited small variability among the different sampling sites, ranging from 33.3% to 40.2% for BAG, 44.9% to 47.0% for BAG, and 34.6% to 43.3%for LFB, with coefficients of variation at 6.0%, 1.6%,and 7.4%, respectively. BGB had the largest mean C content, followed by ABG and LFB (P＜0.05). ABG-,BGB-, and LFB-C stocks ranged from 43 to 109, 23 to 64, and 5 to 20 g/m2, with mean values of 80.82,44.91, and 12.15 g/m2, representing 59%, 33%, and 8% of the total BC stocks, respectively.

3.2 Soil-carbon stocks

SOC content ranged from 0.76 g/kg to 5.73 g/kg,and BD ranged from 1.13 g/cm3to 1.53 g/cm3in the 0–100-cm soil layer at different sites. SOC was highest in the top 5-cm soil layer and decreased gradually with soil depth. In contrast to SOC, BD was lowest in the top 5-cm soil layer and increased gradually with soil depth (Figure 2). Figure 3 showed the SOC-stock distribution in soil profiles for each site.Although there was higher SOC content in the top soil layer, the SOC stock in the top soil layer was lower than that in deep soil layers. The proportion of SOC stock below the 20-cm soil layer ranged from 67% to 76%.

Table 2 Biomass, carbon content, and carbon stocks at nine sampling sites

Figure 2 Vertical distributions of (a) SOC and (b) BD. Standard errors are denoted by horizontal bars; bar having the same letter are not significantly different at α=0.05

Figure 3 Soil-organic-carbon (SOC) stocks in soil profiles for sites at different altitudes;error bars represent standard error

3.3 Changes of carbon stocks along an altitudinal gradient

SOC stocks showed a significant increasing trend with increasing altitude (R2=0.75,P＜0.01). BC stocks also showed a linear increasing trend with altitude (R2=0.53,P＜0.01). The significant relationship between TC stock and altitude was characterized by a linear function of TC=0.003 6 × altitude – 2.49(R2=0.73,P＜0.01) (Figure 4).

Figure 4 Changes in (a) soil-organic-carbon (SOC) stocks, (b) biomass-carbon (BC) stocks,and (c) ecosystem-carbon (EC) stocks along an altitudinal gradient

3.4 Relationships between carbon stock and environmental factors

Figure 5 shows the relationships between SOC and various soil properties in different soil layers. It can be clearly seen that SOC shows a positive correlation with silt content and clay content but a negative correlation with BD, VPG, and sand content. The correlations between SOC and BD, VPG, sand content,and silt content in most soil layers reached a significant level (P＜0.05 orP＜0.01). The significant correlations between SOC and clay content were observed only in the 50–70-cm and 70–100-cm soil layers.

With the exception of BD, VPG, soil-sand content, and silt content, a regression model can be expressed as follows:

SOC=7.92-2.26BD-6.09VPG-0.007sand+0.21 silt (R2=0.50,P＜0.01)

Although soil-C stock tended to decrease with MAT, we identified no significant correlations between biomass-C stocks and MAT. The AGB-C stock increased with increasing MAP (Table 3). In addition, a better fit was obtained between soil- and biomass-C stocks and MAP.Note: * Significant atP≤0.05; ** significant atP≤0.01. AGB, aboveground biomass; BGB, belowground biomass; LFB, litterfall biomass; MAT, mean annual temperature; MAP, mean annual precipitation.

Figure 5 Correlation relationships between SOC and BD (a–f), VPG (g–l), sand content (m–r),and silt content (s–x) at different soil depths

Table 3 Correlations between carbon stocks and climatic variables at nine sampling sites

4 Discussion

On the basis of the field investigation in the Hexi Corridor of China, the results presented here show that the C stocks in desert grasslands have a regular distribution pattern along an altitudinal transect.Across a 250-m altitudinal transect, SOC stocks increased from 2.88 to 3.98 kg/m2and biomass C stocks from 94.3 to 169.2 g/m2. Although variations in C stocks in grassland ecosystems along the altitudinal gradient have been shown in other studies (Zhuanget al., 2007; Leifeldet al., 2009; Djukicet al., 2010), the mechanism responsible for this distribution pattern may differ in desert-grassland ecosystems because of the different environmental conditions.

Direct climate controls over C stocks are well-established (Jobbágy and Jackson, 2000; Austin and Sala, 2002; Kaneet al., 2005), and substantial variations in both temperature and precipitation across the grassland desert have been shown to influence multiple aspects of the grassland C cycle (Burkeet al.,1989; Garcia-Pausaset al., 2007; Songet al., 2012).A number of studies have shown that precipitation determines spatial distribution of biomass C in grasslands (Piaoet al., 2007; Fanet al., 2008; Maet al.,2010), which is further confirmed by the significant positive correlation between MAP and biomass-C stocks in this study. In addition, the study results showed that AGB-C stock negatively correlated with MAT, which is in line with the findings of Yanget al.(2009), who reported that AGB negatively correlated with MAT in arid regions but positively correlated to temperature in wet regions. On the basis of this relation, Yanget al. (2009) concluded that increasing temperature may lead to water loss and, thus, hinder plant growth in grasslands but would promote vegetation growth under humid conditions. From this model,we can postulate that altitude-dependent variations in climate could be the key factor influencing the altitudinal distribution pattern of biomass-C stock identified in this study.

Yanget al. (2008) also reported that soil C in Tibetan alpine grasslands increased with precipitation but was less influenced by temperature. Our study confirmed a significantly positive correlation between MAP and soil-C stocks. However, our study also revealed significantly negative correlations between MAT and soil-C stocks. Increased precipitation has a positive influence on plant growth, thereby increasing the C input into the soil. Further, the relationship between MAT and soil-C stocks can be interrupted by the effect of temperature on the SOC-decomposition processes, increasing C release from the soil (Silveret al., 2000).

Soil is the largest C reservoir in the terrestrial biosphere, containing more C than in vegetation (Yangetal., 2007). In this study, because SOC stocks accounted for more than 95% of the TC stocks, it plays a crucial role in influencing the altitudinal pattern of C stocks. What are the main factors that influence the soil-C-stocks distribution in this desert-grassland ecosystem? Fanget al. (2010) showed that soil-C stock in grasslands was largely determined by soil texture. Our study showed significant correlations between SOC and soil-sand, -silt, and -clay contents (Table 2),which indicated the strong influence of soil texture on variations in the soil-C stocks. In addition, the VPG of the soil in this study exhibited a decreasing trend along the altitudinal gradient (VPG=1,561.89 +e-0.27×altitude,R2=0.55,P＜0.01) (high stone content in a low-altitude area was possibly related to run-off that erodes the topsoil). High stone content in low-altitude areas reduces the proportion of soil in the layers and negatively influences the physical protection of the soil organic matter (Saizet al., 2012), thus resulting in a low soil-C-stock level. Therefore, we can conclude that variations in soil texture could be one of the main factors influencing the altitudinal distribution pattern of soil-C stock in this study.

Other studies have also shown that vegetation composition followed a regular distribution pattern along the altitudinal gradient (Erfanzadehet al.,2013), which is another factor that can influence the ecosystem-C stock (Erfanzadehet al., 2014).However, the effect of vegetation composition on soil-C stock was not observed in this study. This finding could perhaps be because of the lower species richness in the desert grasslands of this study.

We assessed the influence of climate, soil, and vegetation characteristics on C stocks along an altitudinal gradient in a typical desert grassland ecosystem.From the data obtained, we developed a relatively simple function to predict the C stocks in desert grasslands based on the altitude data. The function was shown as follows: TC=0.0036 × altitude-2.49(R2=0.73,P＜0.01). Altitude was chosen as the predictive variable, as it is easily measured, compared to the other environmental variables. By using this function, we predicted the total C stocks of the 1,260-km2grasslands as approximately 4.18 Tg. On average, the ecosystem-C density in the desert grasslands of this study area was estimated as 3.32 kg/m2, which was 64.5% lower than the value (9.34 kg/m2) estimated by Fanget al. (2010) for China's grasslands. The use of the predictive function may allow the estimation of C stocks for larger desert-grassland areas. However, this predictive function will not always be valid because of the strong spatial heterogeneity of soil and climatic conditions when considering the extended spatial scope. Nevertheless, the results of this study could still be used as important field data for C-stock estimation in desert grasslands and could contribute to the development of predictive models of C stocks in arid and semi-arid desert-grassland regions.

5 Conclusions

In summary, the present study suggests that the C stocks in desert grasslands of the Hexi Corridor of China follow an altitudinal pattern, in which both BC and SOC stocks increase with increasing altitude. We propose that variation in the soil texture could be the key factor influencing this distribution pattern. Climate conditions also contributed to the variation of C stocks in this desert-grassland ecosystem. By using a linear-regression function obtained from the present study, approximately 4.18 Tg C was predicted from a 1,260-km2region of desert grasslands.

Acknowledgments:

This study was funded by the Key Research Program of Frontier Sciences of Chinese Academy of Sciences(QYZDJ-SSW-DQC040), the National Key Research and Development Program of China (2017YFC-0504306 and 2017YFC0504304), and the China National Natural Science Foundation (41201284). We thank all staff of the Linze Inland River Basin Research Station for their help in field investigation.We are grateful to the National Meteorological Information Center for providing the meteorological data. We would also like to thank Ali Beamish at the University of British Columbia for her assistance with English language and grammatical editing of the manuscript. We also appreciate the anonymous reviewers for their valuable comments on the manuscript.