Soil organic carbon storage and its influencing factors

Download PDF
0 downloads 0 Views 1MB Size Report
Comment
of soil organic carbon (SOC) storage in riparian woodlands and its ...... Dosskey, M.G., Vidon, P., Gurwick, N.P., Allan, C.J., Duval, T.P., Lowrance, R., 2010.
Catena 153 (2017) 21–29

Contents lists available at ScienceDirect

Catena journal homepage: www.elsevier.com/locate/catena

Soil organic carbon storage and its influencing factors in the riparian woodlands of a Chinese karst area Yunbin Qin, Zhongbao Xin ⁎, Dongmei Wang, Yuling Xiao School of Soil and Water Conservation, Beijing Forestry University, Beijing, China Engineering Research Center of Forestry Ecological Engineering, Ministry of Education (Beijing Forestry University), Beijing, China

a r t i c l e

i n f o

Article history: Received 15 September 2016 Received in revised form 17 January 2017 Accepted 23 January 2017 Available online xxxx Keywords: Riparian Soil organic carbon Karst area Woodland Bamboo

a b s t r a c t Riparian woodlands have recently been recognized as important carbon (C) storage regions with a considerable potential of sequestering C to mitigate global warming. Understanding soil organic carbon (SOC) storage in riparian woodland and the differences in SOC storage between riparian woodlands and other adjacent land types is important to effectively assess SOC storage in riparian woodlands and the influences of riparian land-use changes on SOC pool. Therefore, we thoroughly investigated C storage in the riparian woodlands along the Lijiang River watershed, located in a karst area in southwestern China. The goals of the study were to quantify the SOC density (SOCD) in the riparian woodland (0–20 cm depth of upper soil) [N = 54, the plot area was 1130.88 ± 136.13 m2 (mean value with standard deviation)]; to compare the differences in SOCD in the riparian woodland [bamboodominated (N = 9) and non-bamboo-dominated woodlands (N = 45)], the adjacent grasslands (N = 13) and the farmlands [croplands (N = 17) and orchards (N = 17)]; and to assess the influence of soil texture, plant litter and soil root biomass on the SOCD of the riparian woodlands. The results showed that the average SOCD in the riparian woodlands of the Lijiang River watershed was 35.79 ± 9.51 t/ha. The SOCD in the non-bamboo-dominated woodland was 36.91 ± 9.63 t/ha, which was higher than that in the bamboo-dominated woodlands (29.86 ± 4.90 t/ha) by 7.05 t/ha (about 23.59%) (p b 0.05) and was higher than that in the adjacent grasslands (27.77 ± 7.35 t/ha), croplands (28.93 ± 7.30 t/ha) and orchards (21.26 ± 8.20 t/ha) by 32.91%, 27.58% and 73.61%, respectively (p b 0.01). However, the SOCD in the bamboo-dominated woodlands was only higher than that in the orchard by 8.60 t/ha (about 40.45%) (p = 0.012). In the non-bamboo-dominated woodlands in the Lijiang riparian watershed, the SOCD showed a significant negative correlation with soil sand content (r = −0.69) and a significant positive correlation with the silt (r = 0.59) and clay content (r = 0.61) (p b 0.01). The SOCD showed a significant positive correlation with plant litter (r = 0.44) and soil root biomass (r = 0.38) (p b 0.05). Finally, the results indicate that the non-bamboo-dominated riparian woodland in the karst area stores more SOC compared with the adjacent grassland and farmland. However, converting riparian vegetation into bamboo woodland did not increase the accumulation of SOC and even caused some SOC loss in the Lijiang riparian area. © 2017 Published by Elsevier B.V.

1. Introduction Riparian zones are important corridors between terrestrial and aquatic ecosystems for exchanging material, energy and information and have unique biotic, biophysical and landscape characteristics (Naiman et al., 1993; Vidon et al., 2010; Bedison et al., 2013; Sutfin et al., 2016). As an important part of riparian ecosystems, riparian woodlands play a crucial role in stabilizing riverbanks, trapping and removing nutrients, providing habitats for terrestrial organisms and maintaining ecosystem stability, among other roles, which have been widely and systematically researched (Seth, 1999; Dosskey et al., 2010; Bedison et ⁎ Corresponding author at: School of Soil and Water Conservation, Beijing Forestry University, Beijing, China. E-mail address: [email protected] (Z. Xin).

http://dx.doi.org/10.1016/j.catena.2017.01.031 0341-8162/© 2017 Published by Elsevier B.V.

al., 2013). Riparian woodlands have recently been recognized as important carbon (C) storage areas with a considerable potential to sequester C and mitigate global warming (Hazlett et al., 2005; Cierjacks et al., 2010; Ricker and Lockaby, 2015; Ruffing et al., 2016). Further studies of soil organic carbon (SOC) storage in riparian woodlands and its influencing factors are essential for understanding and enhancing its role in mitigating the increasing atmospheric carbon dioxide (CO2) concentrations. Furthermore, SOC in riparian woodlands is an important measurement index for evaluating soil quality and ecosystem health status (Were et al., 2015; Celentano et al., 2016). Current studies showed that the total riparian area of the earth is 0.8 × 106 to 2.0 × 106 km2, and its SOC storage is approximately 16– 125 Pg, which could account for 0.5–8.0% of the global SOC storage (991–2469 Pg) (Leopold et al., 1964; Tockner and Stanford, 2002; Hiederer and Kochy, 2011; Mitsch and Gosselink, 2015; Sutfin et al.,

22

Y. Qin et al. / Catena 153 (2017) 21–29

2016). Thus, any land use changes in the riparian zones can significantly mitigate or enhance CO2 concentrations in the atmosphere. Changes to the SOC in the riparian zone depend on the ratio between the SOC inflows and outflows during changes to the riparian land use (Guo and Gifford, 2002). Compared to other land uses, woodlands, which are an important part of the riparian zone, usually contain higher SOC (Coleman et al., 2004; Bedison et al., 2013; Ricker et al., 2014). In their research on woodlands and agricultural crop soil in the riparian zone of the North Central United States, Coleman et al. (2004) found that the soil organic carbon density (SOCD) in woodlot woodlands was 175.32 t/ha and was higher than that in agricultural croplands by 112.72 t/ha (0–128 cm depth of upper soil). Bedison et al. (2013) reported that the SOCD in forest riparian sites was higher than that in non-forest riparian sites, with values of 100.3 and 90.6 t/ha (0–30 cm depth of upper soil), respectively. Therefore, the conversion of riparian woodlands to other land uses, particularly to farmlands, can cause a loss of the SOC pool in riparian woodlands (Hazlett et al., 2005; Sutfin et al., 2016). Although many studies have assessed the difference of SOC content in different land use types of riparian zones. However, whether the above results are similar for the soil of riparian woodlands of the karst area is unknown. Karst geomorphology is a unique and important geomorphological type that is generally characterized by extensive outcropping of soluble rocks, rapid and frequent hydrological processes and thin soil layers. It is a vulnerable ecosystem that is created by acidic water on carbonate bedrock. This karst geomorphology has a profound effect on regional water resources and C cycle (Bonacci et al., 2009; Liu et al., 2015). Understanding the differences in SOC content between riparian woodlands and other land uses in the karst region is important to scientifically manage and protect riparian woodlands and can also enhance our understanding of SOC in the riparian zones of this geomorphic region. Due to its location in the ecotone of the river and adjacent upland, the factors influencing SOC accumulation in riparian woodlands are also multiple and complex. The vegetation in the riparian woodland is an important contributor to the SOC cycle via the processes of photosynthesis, root absorption and decomposition, as well as the reduction of plant litter and dead root to produce SOC microcirculation. Vegetation can also stabilize and protect riparian soil and reduce surface erosion and trap sediment, thereby affecting the accumulation and distribution of riparian SOC (Dosskey et al., 2010; Bullinger-Weber et al., 2014; Bätz et al., 2015). The dead plant litter and roots of riparian vegetation are known to be important sources for increasing the SOC storage (Gift et al., 2010; Don et al., 2010; Ricker et al., 2014; Sutfin et al., 2016). Based on the fixed-point observations in the riparian woodland, Ricker et al. (2014) found that the annual average input of leaf litter was 2.4 t C/ha yr in riparian woodlands, and the annual average input of root C was 1.0 t C/ha yr (0–100 cm depth of upper soil). Riparian soil is an important carrier of organic carbon (OC), and its natural properties, such as texture, water content and pH, also have a significant effect on the conversion and decomposition of OC (Bechtold and Naiman, 2006; Rieger et al., 2014; Graf-Rosenfellner et al., 2016). In some studies, the SOC content was positively correlated with soil silt and clay content in riparian soil, i.e., soil with a relatively high fine fraction content (silt and clay) had higher SOC content than soil with coarser fractions (sand) (Bechtold and Naiman, 2006; Hoffmann et al., 2009). To understand the function of riparian woodlands in sequestering SOC, we need to deeply understand the effect of each environmental factor on SOC accumulation. However, few comprehensive studies examining this have been conducted in the riparian zones of karst regions, and further research is needed. To enhance our understanding of SOC in the riparian zones of karst regions and the function of the riparian woodlands in SOC sequestration in this region, the goals of this research were to (Bai et al., 2016) quantify the SOCD of different woodland types in the riparian zone of karst regions, (Bao and Su, 2015) compare the differences of the SOCD in

riparian woodlands and adjacent farmlands and grasslands, and (Bätz et al., 2015) assess the influences of different factors on the SOCD of riparian woodlands. The Lijiang River lies in the southwest region of China and is a karst river where the riparian zone has abundant woodland resources. Therefore, the riparian zone of the Lijiang River watershed, southwestern China was selected as the study area. 2. Materials and methods 2.1. Study area The study area lies in the Lijiang River watershed located in southwestern China (23°23′N–25°59′N, 110°18′E–111°18′E). The Lijiang River belongs to the upper reaches of the Guijiang River in the Pearl River Basin and originates from the northeast side of Mao'er Mountain, which is the highest peak in South China (2141.5 m). This river flows from north to south. The main stream is 214 km long, and the total basin area is 12,285 km2. In 2014, the karst landscape of the Lijiang River watershed was listed as a World Natural Heritage Site. The Lijiang River watershed lies in low latitudes, and is affected by a subtropical moist monsoon climate. The average annual temperature is 17.8–19.1 °C, the annual evaporation is 1377–1857 mm, and the average annual precipitation is 1500–2600 mm (1960–2010) (Duan et al., 2014). This river is a mountain river with a swift, soaring plunge water level that produces strong and frequent riparian scouring. It is mainly recharged by the rain, and the water level changes quickly in response to precipitation. The mean annual total runoff is 41.8 × 109 m3, and the mean annual runoff is 120–130 m3/s (1961–2000). The runoff is extremely uneven throughout a year. The flood season is between March and August, and the runoff nearly accounts for 80% of the total annual runoff. September to February is the dry season. The annual total sediment discharge in the Lijiang River is 124.3 × 105 t, and the mean annual sediment concentration is 0.282 kg/m3 (Huang and Cheng, 2008). The riparian soil of the Lijiang River watershed is dominated by red loam. The upper soil (0–3 m depth) of the riparian zone is composed of sandy loam and silty loam, whereas the lower soil (below 3 m depth) is sandy gravel. The river bed is mainly made up of gravel and coarse sand with little silt. The riparian zone of the Lijiang River watershed has abundant plant resources, with 167 families, 549 genera and 905 species. The major tree species are Pterocarya stenoptera C. DC., Cinnamomumcamphora (L.) J. Presl, Sapiumsebiferum (L.) Roxb. and Bambusa sinospinosa McClure. The main fruit species are Citrus sinensis (L.) Osb., Citrus maxima (Burm.) Merr. and Castaneamollissima Bl. The main shrub species are Flueggeasuffruticosa (Pall.) Baill., Adina rubella Hance and LigustrumsinenseLour. The main herb species are Cynodondactylon (L.) Pers., Polygonumhydropiper L. and Arthraxonhispidus (Thunb.) Makino. Among them, bamboo is a prevailing and local landscape plant, which is widely planted in the riparian of the Lijiang River watershed. It is indispensable to tourism of the Lijiang River and is also an important source of timber economy for the residents living along the river. In the past few years, tourist activities, such as hiking, biking and picnic on the riparian in the Lijiang River watershed have rapidly developed and have seriously affected the riparian woodlands (Wei, 2004; Ren et al., 2014; Bao and Su, 2015) (Fig. 1). 2.2. Sampling selection In this study, 54 sampling plots of riparian woodlands along a 136 km section of in the Lijiang River watershed from Xingan to Yangshuo City were selected to investigate and sample. These plots were based on a field investigation of the riparian zone and regional remote sensing images of the Lijiang River watershed. To standardize the sampling, all the plots were parallel to the river; the width of the longitudinal parallel to the river was not b30 m, and the width of the transverse was not b8 m. The onset transverse boundary of the plot was

Y. Qin et al. / Catena 153 (2017) 21–29

23

Fig. 1. Location of the sampling sites along the Lijiang River watershed.

started from the place underneath of the longest tree canopy near to the river channel. If the tree canopy boundary exceeds the riparian, the onset boundary of the plot transverse was started from the river bank. The sampling area was not b1000 m2 (the mean area of the sample was 1130.88 ± 136.13 m2; the smallest was 1000 m2, and the largest was 1910.33 m2), and spatially separated plots were located at least 1 km from one another to avoid pseudo-replication. If the uplands adjacent to the riparian woodlands were farmlands or the lowlands between the riparian woodland and the river were grasslands, a sampling plot was included in this farmland or grassland. The area of the sampling plot in the farmland and grassland was similar in size to the adjacent riparian woodland. The onset transverse boundary of the adjacent farmlands was closed to the end boundary of the riparian, so the farmlands were hard to be submerged by the flood. The onset transverse boundary of the adjacent grasslands was started from the river bank and the end transverse boundary was bordered on the onset boundary of the woodlands. Therefore, inundation frequency of the grasslands was higher than that in the adjacent woodlands. Finally, the total number of sampling plots in this study was 101, including 54 in riparian woodlands, 34 in adjacent farmlands (orchards: N = 17; croplands: N = 17) and 13 in adjacent grasslands. 2.3. Plot investigation and sample collection analysis 2.3.1. Plot investigation The investigation and the collection were undertaken between April and June 2015. The information surveyed and recorded from the sampling plot included geographical position, elevation, terrain (slope and the length of slope), plant composition and disturbance from human

activities. When investigating the plants of the riparian woodlands, the tree and shrub species were used in a major quadrat survey, i.e., all trees and shrubs in the sampling plot were examined and recorded (Burton and Samuelson, 2008). The recorded information included the species name, height, and diameter at breast height (DBH) or basal diameter. When measuring the DBH, the individuals with a diameter ≥ 2.5 cm were measured, and the DBH was recorded; individuals with a diameter b 2.5 cm were only measured and recorded at the base. The woody bamboo in the sampling plot usually grew and formed a clump with dense stands. Therefore, we recorded the number of stems of each clump and the diameter of each stem. The basal area of each clump was counted by the basal area of all stems, and the numbers of the individuals in each clump were recorded as 1 at counting the density. The investigation of herb species in the sampling plot was used a method of small quadrat survey, i.e., three 1 m × 1 m quadrats were randomly selected for study (Vockenhuber et al., 2011). The recorded information included the species name, species density, average species height and quadrat coverage. The information recorded from the farmland sampling plot included the distance to the adjacent riparian woodlands, farmland type, crop type, irrigation and fertilization situation and the degree of inundation in flood seasons. The recorded information in the grassland sampling plot included the distance to the adjacent riparian woodlands, plant species, human activities and the degree of inundation in flood seasons. 2.3.2. Soil sampling and analysis The riparian soil layer in the Lijiang River watershed was thin, and the upper soil was frequently affected by the flooding. Therefore, the depth of the soil sample in this study was 0–20 cm depth of the upper

24

Y. Qin et al. / Catena 153 (2017) 21–29

soil. In each sampling plot, three soil samples were randomly collected using a stainless steel cutting ring (5.0 cm high and 5.0 cm in diameter) to measure the soil bulk density (BD) of the topsoil. Five random soil samples (0–20 cm depth of upper soil) were collected using a soil auger that was 20 cm long. The five soil samples were then manually homogenized to form a composite sample, and one quarter of this sample was transported to laboratory for analysis. In addition, we randomly selected three 1 m2 quadrats (1 m × 1 m) to collect plant litter in the sampling plot. If the sampling plot had no plant litter, the plant litter weight was recorded as 0. Three soil root samples were collected from 100 cm2 (10 cm × 10 cm) quadrants that were randomly selected in the sampling plot. The collected depth of soil root was 0–20 cm of upper soil. Then these soil samples were passed through a 0.25 mm sieve by using river water. And the retained soil root was collected and brought back to the laboratory. After being air-dried, impurities and roots were removed from the soil samples collected from the field, which were then weighed. And all the soil samples were passed through a 2-mm sieve, and the gravel that could not pass through this sieve size was weighed. One part of the subsample was analyzed to determine soil texture; the other subsample was completely passed a 0.25-mm sieve to determine the SOC content. The soil texture was determined by the hydrometer method, and classified per the ISSS (the International Soil Science Society) methodology (Bouyoucos, 1962; Gee and Or, 2002). The weight of plant litter and root biomass was measured after drying to a constant weight at 60 °C. The SOC content was measured by using the K2Cr2O7H2SO4Walkley-Black oxidation method (Nelson and Sommers, 1982). 2.3.3. Data analysis The division of riparian woodland types: the riparian woodland in this study area was divided into bamboo-dominated and non-bamboo-dominated woodlands according to the different dominant trees in the woodlands. If the proportion of the density or basal area of the mature bamboo (DBH ≥ 10 cm) accounted for 80% of all mature trees in the riparian woodland sampling plot, this plot was regarded as bamboo-dominated woodland. Based on the plant height and diameter, the plants in the sampling plot were divided into four size classes: (Bai et al., 2016) overstory canopy (DBH ≥ 10 cm), (Bao and Su, 2015) middle-story canopy (2.5 cm ≤ DBH b 10 cm), (Bätz et al., 2015) understory canopy (DBH b 2.5 cm) and (Bechtold and Naiman, 2006) herb layer. The investigated shrubs were also divided into corresponding diameter size classes to calculate the diversity. In addition, the upland farmlands were divided into orchards and croplands according to the different crops in the farmlands. Calculation of the formulas of the ecological indicators and the SOCD: The richness index was calculated using the following equation (Margalef, 1958): R ¼ ðS−1Þ= ln ðNÞ

ð1Þ

where, R is the species richness index, S is the total number of species, and N is the total number of individuals of each species. The Shannon Wiener Diversity index (H′) was computed using the following equation (Shannon and Weaver, 1963): H0 ¼ −Σ Pi ln ðPi Þ

ð2Þ

where, Pi = ni/N (ni = the number of individuals of a species, and N = the total number of individuals of each species). The SOC content was computed using the equation (Guo and Gifford, 2002): SOC ¼ SOM  0:58

ð3Þ

where, SOC is the soil organic carbon content (%), and SOM is the soil organic matter (%).

The SOCD was computed using the equation (Guo and Gifford, 2002): SOCD ¼ SOC  BD  D

ð4Þ

where, SOCD is the soil organic carbon density (t/ha), SOC is the soil organic carbon content (%), D is the soil-layer depth (cm), and BD is the soil bulk density (g/cm3). Statistical analyses were performed in Excel 2010 and SPSS 18.0 (Statistical Package for the Social Sciences). Spearman's correlation was used to analyze the correlation between the SOCD in the Lijiang riparian zone and the influencing factors (i.e., soil texture, plant litter and soil root biomass). Analysis of variance (ANOVA) was performed to determine the significant differences among the SOCD and other soil variables. Fisher's LSD test was used to compare the mean values among SOCD and other soil variables when the results of ANOVA were significant at p ≤ 0.05. An independent-sample t-test was performed to determine significant differences of investigated plant variables between bamboodominated woodlands and non-bamboo-dominated woodlands. Finally, the SOCD and all factors (soil texture, plant litter and soil root biomass) in the riparian woodland were normalized, and linear multiple regression models with backward and enter selection techniques were used to analyze the influence of the factors on SOCD and identify the main predictor variables. The SOCD was the dependent variable, and soil texture, plant litter and soil root biomass were the independent variables. At the backward removal procedure, a significance test was performed to determine which independent variables should be included in the model (F ratio to enter, 0.05; F ratio to remove, 0.1). All the statistical analyses were performed with a significance level of p ≤ 0.05. 3. Results 3.1. Plant inventory of riparian woodlands A total of 260 plant species was recorded in the riparian woodlands of the Lijiang River watershed, where the number of tree (DBH ≥ 10 cm), shrub and herb species was 41, 72 and 147, respectively. The dominant tree species were Pterocarya stenoptera, bamboo, camphor tree and Chinese tallow tree in the overstory canopy. The density of all the tree and shrub species with diameters ≥ 2.5 cm was 959.04 ± 447.87 individuals/ha in the riparian woodland, where the average basal area of these plants was 25.74 ± 10.65 m2/ha. In the middle-story canopy, the respective basal areas and the densities of the trees and shrubs in the bamboo-dominated woodland were 0.14 ± 0.08 m2/ha and 109.86 ± 86.98 individuals/ha and were lower than those in the nonbamboo-dominated woodlands by 0.74 m2/ha and 428.37 individuals/ ha, respectively (p b 0.01). The average richness and Shannon diversity indexes of the riparian woodland were 3.34 ± 1.24 and 2.23 ± 0.47, respectively. In the bamboo-dominated woodland, the respective richness of the overstory and the middle-story canopies was 0.34 ± 0.10 and 0.86 ± 0.58, which were lower than that in the non-bamboo-dominated woodland by 0.64 and 0.65, respectively (p b 0.05). The respective Shannon diversity indexes of the overstory and the middle-story canopies were 0.47 ± 0.28 and 0.72 ± 0.48, which were lower than that in the nonbamboo-dominated woodland by 0.35 and 0.53, respectively (p b 0.05). In the herb layer, the bamboo-dominated woodland had 79.86% less the plant cover (16.00 ± 15.57%) than the non-bamboodominated woodland (p b 0.01) (Table 1). Thus, the bamboo-dominated woodland, which is very common in the riparian zone of the Lijiang River watershed, is a special woodland type with obvious regional features with significant differences in the plant community compared to the non-bamboo-dominated woodlands.

Y. Qin et al. / Catena 153 (2017) 21–29

25

Table 1 Attributes of the plants in the riparian woodland (mean values with S.D. in parentheses). Basal area (m2 ha−1)

Density (individuals ha−1)

24.80 (9.73)a

0.88 (0.67)a

494.84 538.23 (230.85)a (381.15)a

4022.45 (2953.08)a

0.98 (0.69)a

1.51 (0.65)a

3.13 (1.22)a

0.82 (0.55)a

1.25 (0.44)a

2.09 (0.46)a

78.75 (12.75)a

25.90 (15.65)a

0.14 (0.08)b

464.03 109.86 (197.44)a (86.98)b

2165.92 (2304.64)a

0.34 (0.10)b

0.86 (0.58)b

2.72 (1.52)a

0.47 (0.28)b

0.72 (0.48)b

1.79 (0.96)a

16.00 (15.57)b

Richness index Shannon index Coverage of herb Overstory Middle-story Overstory Middle-story Understory Overstory Middle-story Understory Overstory Middle-story Understory layer (%) canopy canopy canopy canopy canopy canopy canopy canopy canopy canopy canopy Non-bamboo dominated woodlands (N = 45) Bamboo dominated woodlands (N = 9)

Within a row, mean values followed by the same letter do not significantly differ at the p b 0.05 level.

3.2. Differences in soil texture, root biomass and plant litter The soil texture (d b 2 mm) of the riparian woodlands in the Lijiang River watershed was mainly comprised of sand. The average sand, silt and clay content were 61.05 ± 13.43%, 20.99 ± 9.53% and 17.97 ± 6.91%, respectively. A one-way ANOVA indicated that land-use had significant effect on the soil sand (F = 4.46), silt (F = 17.57) and clay content (F = 8.09) (p b 0.01). The non-bamboo-dominated woodland had 52.94%, 53.98% and 66.67% more clay content (18.21 ± 7.24%) than the grasslands, orchards and croplands, respectively (p b 0.01); and had 46.17% and 50.24% less silt content (19.91 ± 7.98%) than the orchards and croplands, respectively (p b 0.01). The bamboo-dominated woodland had 33.24% and 27.79% less silt content (26.71 ± 14.77%) than the croplands and orchards, respectively (p b 0.05), and had 40.08%, 52.66% and 41.03% more clay content (16.67 ± 4.93%) compared to the grasslands, croplands and orchards, respectively (p b 0.05). A one-way ANOVA indicated that land-use had a significant effect on the soil root biomass (F = 29.04, p b 0.01). The average soil root biomass at 0–20 cm depth was 410.30 ± 229.37 g/m2 in the riparian woodland, whereas the soil root biomass was 557.11 ± 261.03 g/m2 in the bamboo-dominated woodland, which was higher than that in the non-bamboo-dominated woodlands (340.76 ± 180.97 g/m2) by 216.35 g/m2 (p b 0.01). The soil root biomass in the non-bamboo-dominated woodlands was significantly higher than that in the orchards and croplands by 223.07 g/m2 and 282.68 g/m2 (approximately 1.89 and 4.88 times), respectively (p b 0.01). All the riparian woodland sampling plots had plant litter on the surface soil, whereas the frequency of the occurrence of plant litter in the grasslands, orchards and cropland sites was 23.08%, 35.29% and 0%, respectively. A one-way ANOVA indicated that land-use had a significant effect on the plant litter content (F = 40.19, p b 0.01). The average plant litter content was 0.31 ± 0.22 kg/m2 in the riparian woodlands. In the bamboo-dominated woodland, the plant litter content was

0.52 ± 0.22 kg/m2, which was higher than that in the non-bamboodominated woodland by 0.31 kg/m2 (about 147.62%) (p b 0.01). The non-bamboo-dominated woodland had approximately 11 and 7 times more plant litter content (0.21 ± 0.15 kg/m2) than the grasslands and orchards, respectively (p b 0.01) (Table 2). 3.3. SOC storage in the different riparian land use types The one-way ANOVA indicated that land-use type had a significant effect on the SOCD (F = 11.92, p b 0.01). The average SOCD in the riparian woodlands along the Lijiang River watershed was 35.79 ± 9.51 t/ha, whereas the SOCD in the non-bamboo-dominated woodland was 36.91 ± 9.63 t/ha, which was higher than that in the bamboo-dominated woodland (29.86 ± 4.90 t/ha) by 7.05 t/ha (about 23.59%) (p b 0.05). The SOCD in the grasslands, croplands and orchards were 27.77 ± 7.35 t/ha, 28.93 ± 7.30 t/ha and 21.26 ± 8.20 t/ha, respectively. The non-bamboo-dominated woodland had 32.91%, 27.58% and 73.61% more SOCD than the grasslands, croplands and orchards, respectively (p b 0.01), whereas the bamboo-dominated woodland had only 40.45% more SOCD compared to the orchards (p = 0.012). The orchard had the smallest SOCD, which was lower than that in the riparian woodlands, croplands and grasslands by 40.61%, 26.52% and 30.66%, respectively (p b 0.05) (Fig. 2). 3.4. Relationships between the SOCD and environmental factors in the riparian woodlands The SOCD in the riparian non-bamboo-dominated woodland showed a significantly negative correlation with the soil sand content (r = −0.69, p b 0.01), a significantly positive correlation with the soil silt (r = 0.59) and clay content (r = 0.61) (p b 0.01), and a significantly positive correlation with the plant litter (r = 0.38) and soil root biomass (r = 0.44) (p b 0.05). However, no significant correlations were found

Table 2 Attributes of environmental factors in riparian woodlands, adjacent grasslands and farmlands (mean values with S.D. in parentheses). Riparian woodlands (N = 54)

Sand content (%) Silt content (%) Clay content (%) Soil root biomass (g/m2) Frequency of the presence of litter (%) Plant litter (kg/m2)

Total woodlands (N = 54)

Non-bamboo dominated woodlands (N = 45)

Bamboo dominated woodlands (N = 9)

61.05 (13.43) 20.99 (9.53) 17.97 (6.91) 376.82 (197.87) 100 0.26 (0.20)

61.88 (13.34)a 19.91 (7.98)a 18.21 (7.24)a 340.76 (163.69)a 100 0.21 (0.15)a

56.62 (13.80)ab 26.71 (14.77)a 16.67 (4.93)a 557.11 (261.03)b 100 0.52 (0.22)b

Within a row, mean values followed by the same letter do not significantly differ at the p b 0.05 level.

Adjacent grasslands (N = 13)

Adjacent farmlands (N = 34) Total

Orchards (N = 17)

Croplands (N = 17)

62.82 (12.42) ac 25.28 (10.72)a 11.90 (3.11)b 576.30 (275.78)b 23.08 0.017 (0.036)c

50.13 (14.06) 38.50 (11.41) 11.37 (4.39) 87.88 (65.61) 17.65 0.013 (0.041)

51.19 (16.11)bc 36.99 (12.33)b 11.82 (5.26)b 117.69 (84.59)c 35.29 0.025 (0.055)c

49.07 (12.09)b 40.01 (10.58)b 10.92 (3.41)b 58.08 (48.62)d 0 0d

26

Y. Qin et al. / Catena 153 (2017) 21–29

4. Discussion 4.1. Comparison of carbon storage among the Lijiang riparian watershed and other regions

Fig. 2. SOCD in different land use types in the riparian zone (RW: riparian woodlands; BW: bamboo-dominated woodlands; NBW: non-bamboo-dominated woodlands; AG: adjacent grasslands; AF: adjacent farmlands; AO: adjacent orchards; AC: adjacent croplands. The numbers in the parentheses are the numbers of the sampling sites. Error bars represent 95% confidence interval in all panels. The different letters indicate significant differences at p b 0.05. Lower-case letters show the SOCD differences in BW, NBW, AG, AO and AC; capital letters show the SOCD differences in RW, AG and AC).

between the SOCD and soil texture, soil root biomass and plant litter in the bamboo-dominated woodland (Table 3). Therefore, this study used only the SOCD and all the environmental factors (soil texture, plant litter and soil root biomass) in the non-bamboo-dominated woodland to construct multiple regression models to analyze the contributions of all the factors to the SOCD variation. In addition, obvious collinearity existed between the soil sand, silt and clay contents. Thus, only the soil clay content, which had a stronger correlation with the SOCD than the sand and silt content, was used to represent the soil texture to construct the multiple regression models. The results were as follows:  SOCD ¼ 33:32 þ 12:38  ðsoil clay contentÞ r 2 ¼ 0:36; p ¼ 0:014 ð5Þ

SOCD ¼ 30:23 þ 11:64  ðsoil clay contentÞ þ 5:66  ðplant litterÞ  ð6Þ þ5:73  ðsoil root biomassÞ r2 ¼ 0:47; p ¼ 0:046

In the riparian non-bamboo-dominated woodland, the soil clay content was the best predictor that explained the SOCD variation, which could account for 36.00% of the SOCD variation (Eq. (5)). The combined soil clay content, plant litter and soil root biomass accounted for 47.00% of the SOCD variation (Eq. (6)).

When differences in sampling depth were considered, the SOCD in the riparian woodland of the Lijiang River watershed (35.79 ± 9.51 t/ha, 0–20 cm depth of upper soil) was close to that measured in the riparian woodland of other regions. For example, the SOCD was 33 t/ha (0–27 cm depth of upper soil) in the riparian forest of the boreal lakes in northeastern Ontario (Hazlett et al., 2005) and 34–48 t/ha (0– 25 cm depth of upper soil) in a floodplain forest of the southeastern Atlantic Coastal Plain, USA (Ricker and Lockaby, 2015). However, the SOCD in the riparian woodland of the Lijiang River watershed was lower than that in the riparian forests of the Danubian floodplains, where the SOCD was 41–58 t/ha (depth of Ah horizon: 15–20 cm; Cierjacks et al., 2010) and was far lower than that reported by Bedison et al. (2013), i.e., 100.3 and 90.6 t/ha (0–30 cm depth of upper soil) for the forested and non-forested riparian zones, respectively, of the Atlantic Coastal Plain of the Delaware River Basin. These differences were partly related to the riparian woodland of the Cierjacks and Bedison's study having a higher soil organic horizon and soil clay content than the riparian woodlands of the Lijiang River watershed. The average soil clay content was 21.4% in the riparian zone of the Bedison's study compared to 17.97% at the Lijiang River watershed. Several studies had shown that fine fractions (clay and silt) can retain more soil organic matter than the coarser fractions (Bechtold and Naiman, 2006; Hoffmann et al., 2009; Bullinger-Weber et al., 2014). In addition, the Lijiang River is a mountain river with a swift, soaring plunge water level that produces strong and frequent riparian scouring. The high precipitation and temperature also promote the decomposition of SOC in this region. The human disturbances, such as hiking and picnic, to the riparian environment in this watershed also became increasingly frequent as a result of tourism development. Thus, these factors collectively contribute to the riparian zone of the Lijiang River watershed characterized by low SOC contents compared to the riparian of Danubian floodplains and Atlantic Coastal Plain. 4.2. The SOCD in the riparian non-bamboo-dominated woodland of the Lijiang River watershed Land use changes can increase SOC storage and cause SOC storage loss, depending on the ratio between the SOC inflows and outflows (Guo and Gifford, 2002). In the Lijiang River watershed, the riparian non-bamboo-dominated woodland had a higher SOCD than the adjacent grasslands and farmlands (p b 0.05) (Fig. 2), implying that the conversion of riparian non-bamboo dominated woodlands to grasslands and farmlands, and particularly to orchards, could cause a loss of the SOC pool in non-bamboo-dominated riparian woodlands. This conversion could lead to a 27%–73% SOC storage loss in the riparian woodland of the Lijiang River watershed. In other riparian regions, differences in the SOC content among the different riparian land use types were also found (Coleman et al., 2004; Bedison et al., 2013; Ricker et al., 2014).

Table 3 Relationships of the SOCD and environmental factors in the riparian zone. Riparian woodlands (N = 54)

Sand content (%) Silt content (%) Clay content (%) Soil root biomass (g/m2) Plant litter (kg/m2)

Total woodlands (N = 54)

Non-bamboo dominated woodlands (N = 45)

Bamboo dominated woodlands (N = 9)

−0.64⁎⁎ 0.55⁎⁎ 0.57⁎⁎ −0.15 −0.018

−0.69⁎⁎ 0.59⁎⁎ 0.61⁎⁎ 0.38⁎ 0.44⁎

−0.15 0.18 0.11 −0.12 0.10

⁎ Indicates a significant difference at the 0.05 level. ⁎⁎ Indicates a significant difference at the 0.01 level.

Adjacent grasslands (N = 13)

Adjacent farmlands (N = 34)

Total samples (N = 101)

−0.78⁎ 0.69⁎ 0.74⁎ 0.48⁎ 0.12

−0.52⁎⁎ 0.44⁎ 0.49⁎⁎ −0.34 −0.30

−0.25⁎ −0.13 0.66⁎⁎ 0.19 0.31⁎⁎

Y. Qin et al. / Catena 153 (2017) 21–29

These differences were due partly to the soil of riparian non-bamboodominated woodlands near the Lijiang River watershed having higher soil clay content. This study showed a significantly positive correlation between the SOCD and the soil clay content in the riparian non-bamboo-dominated woodland (r = 0.61, p b 0.01) (Table 3). For example, the soil with a relatively high clay content had a higher SOCD than that with a low clay content. The soil clay content was 18.21% in the riparian non-bamboo-dominated woodland, which was higher than that in the grasslands, orchards and croplands by 52.94%, 53.98% and 66.67%, respectively (p b 0.01) (Table 2). The high plant litter and soil root biomass of the riparian non-bamboo-dominated woodland also contributed to the riparian woodland with a higher SOCD than the adjacent farmlands and grasslands. The SOCD in the riparian non-bamboo-dominated woodland showed a significantly positive correlation with the plant litter and soil root biomass, with an r value of 0.44 and 0.38 (p b 0.05), respectively (Table 3). This suggests that the greater the plant litter and soil root biomass in the riparian woodland, the higher the SOC content in the riparian soil. Many investigations have confirmed that plant litter and soil root biomass were the main sources of SOC accumulation (Don et al., 2010; Rieger et al., 2014; Sutfin et al., 2016). The plant litter content in the riparian non-bamboo-dominated woodland was higher than that in the grasslands and orchards by 11 and 7 times, respectively (p b 0.01). The soil root biomass in the riparian non-bamboo-dominated woodland was higher than that in the orchards and croplands by 1.89 and 4.88 times, respectively (p b 0.01) (Table 2). Thus, the SOC source in the riparian non-bamboo-dominated woodland was more than that in the grasslands and farmlands. Because this sampling time was before the flood season, the plant litter remained in the riparian woodland. However, during the flood season, surface soil of some low riparian woodland regions is vulnerable to erosion and deposition from the periodic floods, which can wash away, bury and deposit riparian litter and thus affect the local and regional accumulation of SOC (Ellis et al., 1999; González et al., 2010; Drouin et al., 2011; Rieger et al., 2014). The influence of plant litter on riparian SOC storage needs to be further explored. The intercepted slope runoff in the riparian woodland can also promote the accumulation of SOC in the riparian non-bamboo-dominated woodland. Due to its location below the upland farmlands, the plants and soil of the riparian woodland can effectively intercept nutrients and sediments carried by the slope runoff flowing from the upland farmlands, which enriches the soil nutrients of the riparian woodland (Jose, 2009; Yuan et al., 2009; Fortier et al., 2016). The nutrients flowing from the farmlands are mainly nitrogen and phosphorus, and some studies have shown that soil nitrogen can also promote the accumulation of SOC (Luo et al., 2004; Morris et al., 2007). The grasslands lie at the base of the riparian zone, where flood erosion is more frequent than in the adjacent riparian woodland. This flood erosion may offset the effect of nutrient enrichment to some extent.

27

4.3. SOCD in the riparian bamboo-dominated woodland The SOCD in the riparian bamboo-dominated woodland was 29.86 ± 4.90 t/ha, which was smaller than that in the non-bamboodominated woodland by 7.05 t/ha (p b 0.05). No significant differences were found between the adjacent farmlands and grasslands (Fig. 2), although the soil clay and plant litter content of the riparian bamboodominated woodland were significantly higher than in the adjacent farmlands and grasslands, and the soil root biomass content of the riparian bamboo-dominated woodland was higher than that in the farmlands (p b 0.05) (Table 3). Some previous studies have also found that the non-bamboo dominated woodlands with a higher SOC content are comparable to the bamboo woodlands in other regions (Guan et al., 2015; Bai et al., 2016). From the soil investigation in the secondary forest and the Moso bamboo converted from the secondary forest in Shitai County, Anhui Province, China, Guan et al. (2015) found that SOC storage was significantly smaller than that in the secondary forest. The SOCD in the Moso bamboo was 118.25 t/ha (0–50 cm depth of upper soil), which was lower than that in the secondary forest by 41.94% (p b 0.05). Soil erosion is regarded as a main reason causing the soil with a low SOC content (Schilling et al., 2009; Garzon-Garcia et al., 2015). In the bamboo-dominated woodland, the seed germination and seedling growth of other plants are usually limited by the growth of bamboo, such as the relatively dense distribution of mature bamboo, high bamboo litter cover during the non-flood season and root competition, which can cause the plant deficiency of the understory canopy, resulting in the bare surface soil and the loss of SOC (Griscom and Ashton, 2006; Rother et al., 2009; Budke et al., 2010; Larpkern et al., 2011). The plant coverage of the middle-story canopy and herb layer in the riparian bamboo-dominated woodland was significantly lower than that in the non-bamboo woodland of the Lijiang River watershed (Table 1). In this study, most of the sampling plots in the bamboo-dominated woodlands are in a relatively low landscape position that is prone to frequent flood disruptions in the flood season. Due to the lack of protection of the surface soil, the riparian bamboo-dominated woodlands easily lose their surface SOC from the scouring and erosion of floods and rains. In addition, bamboo is one of the fastest growing plants, with a growth rate ranging from 30 to 100 cm per day in the growing season. A stem can reach its full height in 2– 3 months (Zhou et al., 2005). The fast growth consumes a lot of SOC, whereas the SOC pool cannot be replenished due to the plant litter carried away by the flood and the stem cut by the farmer, which can cause the SOC content decline. As an excellent and local landscape tree in the riparian of the Lijiang River watershed, bamboo is indispensable to tourism in the Lijiang River watershed and is also an important source of timber economy for the riparian residents (Fig. 3). Thus, how to construct a bamboo ecosystem in harmony with the landscape and ecology is very important for future riparian bamboo planting management.

Fig. 3. Riparian bamboo-dominated woodlands of the Lijiang River watershed, (a) bamboo is an important landscape of the Lijiang River; (b) lack of plant cover in the bamboo-dominated woodlands.

28

Y. Qin et al. / Catena 153 (2017) 21–29

4.4. SOCD in the orchard

References

In the study area, the SOCD in the orchard was the lowest at 21.26 ± 8.20 t/ha, which was lower than that in the woodland, cropland and grassland by 40.61%, 26.52% and 30.66%, respectively (p b 0.05) (Fig. 2). Analysis suggests that the management mode of the orchard is the main reason for its low SOCD. In the Lijiang River watershed orchards, farmers take away fruit, trim the branches, and remove the litter, which decreases the organic compounds entering the soil. Clearing weeds under the trees and plowing the soil in the orchards also causes stronger soil erosion. However, croplands only lack plant cover in winter when precipitation is relatively low and soil erosion is also relatively weak. Thus, the existing management modes need to be changed and new orchard management models need to be established to reduce soil erosion and increase SOC sequestration. Some studies have found that orchards managed with conservation practices, such as growing grass and leguminous cover crops and mulching the ground, can effectively protect the surface soil from erosion (Umali et al., 2012; Guimarães et al., 2013).

Bai, S.B., Conant, R.T., Zhou, G.M., Wang, Y.X., Wang, N., Li, Y.H., Zhang, K.Q., 2016. Effects of moso bamboo encroachment into native, broad-leaved forests on soil carbon and nitrogen pools. Sci. Report. 6, 31480. Bao, Q.Q., Su, W.B., 2015. Study on ecosystem health evaluation of river scenic spot: taking Lijiang River scenic spot as an example. J. Nat. Dis. Sci. 24, 122–127 (in Chinese with English abstract). Bätz, N., Verrecchia, P.E., Lane, S.N., 2015. Organic matter processing and soil evolution in a braided river system. Catena 126, 86–97. Bechtold, J.S., Naiman, R.J., 2006. Soil texture and nitrogen mineralization potential across a riparian toposequence in a semi-arid savanna. Soil Biol. Biochem. 38, 1325–1333. Bedison, J.E., Scatena, F.N., Mead, J.V., 2013. Influences on the spatial pattern of soil carbon and nitrogen in forested and non-forested riparian zones in the Atlantic Coastal Plain of the Delaware River Basin. For. Ecol. Manag. 302, 200–209. Bonacci, O., Pipan, T., Culver, D.C., 2009. A framework for karst ecohydrology. Environ. Geol. 56, 891–900. Bouyoucos, G.J., 1962. Hydrometer method improved for making particle size analyses of soils. Agron. J. 54, 464–465. Budke, J.C., Alberti, M.S., Zanardi, C., Baratto, C., Zanin, E.M., 2010. Bamboo dieback and tree regeneration responses in a subtropical forest of South America. For. Ecol. Manag. 260, 1345–1349. Bullinger-Weber, G., Le Bayon, R.C., Thébault, A., Schlaepfer, R., Guenat, C., 2014. Carbon storage and soil organic matter stabilisation in near-natural, restored and embanked Swiss floodplains. Geoderma 228–229, 122–131. Burton, M.L., Samuelson, L.J., 2008. Influence of urbanization on riparian forest diversity and structure in the Georgia Piedmont, US. Plant Ecol. 195, 99–115. Celentano, D., Rousseau, G.X., Engel, V.L., Zelarayan, M., Oliveira, E.C., Araujo, A.C.M., de Moura, E.G., 2016. Degradation of riparian forest affects soil properties and ecosystem services provision in Eastern Amazon of Brazil. Land Degrad. Dev. Cierjacks, A., Kleinschmit, B., Babinsky, M., Kleinschroth, F., Markert, A., Menzel, M., Ziechmann, U., Schiller, T., Graf, M., Lang, F., 2010. Carbon stocks of soil and vegetation on Danubian floodplains. J. Plant Nutr. Soil Sci. 173, 644–653. Coleman, M.D., Isebrands, J.G., Tolsted, D.N., Tolbert, V.R., 2004. Comparing soil carbon of short rotation poplar plantations with agricultural crops and woodlots in North Central United States. Environ. Manag. 33, S299–S308. Don, A., Schumacher, J., Freibauer, A., 2010. Impact of tropical land-use change on soil organic carbon stocks—a meta-analysis. Glob. Chang. Biol. 17, 1658–1670. Dosskey, M.G., Vidon, P., Gurwick, N.P., Allan, C.J., Duval, T.P., Lowrance, R., 2010. The role of riparian vegetation in protecting and improving chemical water quality in streams. J. Am. Water Resour. Assoc. 46, 261–277. Drouin, A., Saint-Laurent, D., Lavoie, L., Ouellet, C., 2011. High-precision elevation model to evaluate the spatial distribution of soil organic carbon in active floodplains. Wetlands 31, 1151–1164. Duan, W.J., Wang, J.Y., Zhang, L.J., Li, H.F., Huang, H.Q., 2014. Characteristics of precipitation in Lijiang River basin during 1960–2010. J. Hydrol. 34, 88–93 (in Chinese with English abstract). Ellis, L.M., Molles Jr., M.C., Crawford, C.S., 1999. Influence of experimental flooding on litter dynamics in a Rio Grande riparian forest, New Mexico. Restor. Ecol. 7, 193–204. Fortier, J., Truax, B., Gagnon, D., Lambert, F., 2016. Potential for hybrid poplar riparian buffers to provide ecosystem services in three watersheds with contrasting agricultural land use. Forests 7, 37. Garzon-Garcia, A., Olley, J.M., Bunn, S.E., 2015. Controls on carbon and nitrogen export in an eroding catchment of south-eastern Queensland, Australia. Hydrol. Process. 29, 739–751. Gee, G.W., Or, D., 2002. 2.4 Particle-size analysis. Methods of Soil Analysis: Part 4 Physical Methods, pp. 255–293. Gift, D.M., Groffman, P.M., Kaushal, S.S., Mayer, P.M., 2010. Denitrification potential, root biomass, and organic matter in degraded and restored urban riparian zones. Restor. Ecol. 18, 113–120. González, E., Muller, E., Gallardo, B., Comín, F.A., González-Sanchis, M., 2010. Factors controlling litter production in a large Mediterranean river floodplain forest. Can. J. For. Res. 40, 1698–1709. Graf-Rosenfellner, M., Cierjacks, A., Kleinschmit, B., Lang, F., 2016. Soil formation and its implications for stabilization of soil organic matter in the riparian zone. Catena 139, 9–18. Griscom, B.W., Ashton, P.M.S., 2006. A self-perpetuating bamboo disturbance cycle in a neotropical forest. J. Trop. Ecol. 22, 587–597. Guan, F.Y., Tang, X.L., Fan, S.H., Zhao, J.C., Peng, C., 2015. Changes in soil carbon and nitrogen stocks followed the conversion from secondary forest to Chinese fir and moso bamboo plantations. Catena 133, 455–460. Guimarães, D.V., Gonzaga, M.I.S., Silva, T.O., Silva, T.L., Dias, N.S., Matias, M.I.S., 2013. Soil organic matter pools and carbon fractions in soil under different land uses. Soil Tillage Res. 126, 177–182. Guo, L.B., Gifford, R.M., 2002. Soil carbon stocks and land-use change: a metaanalysis. Glob. Chang. Biol. 8, 345–360. Hazlett, P.W., Gordon, A.M., Sibley, P.K., Buttle, J.M., 2005. Stand carbon stocks and soil carbon and nitrogen storage for riparian and upland forests of boreal lakes in northeastern Ontario. For. Ecol. Manag. 219, 56–68. Hiederer, R., Kochy, M., 2011. Global Soil Organic Carbon Estimates and the Harmonized World Soil Database, JRC Scientific and Technical Report. European Union European Joint Commission Centre, Luxembourg. Hoffmann, T., Glatzel, S., Dikau, R., 2009. A carbon storage perspective on alluvial sediment storage in the Rhine catchment. Geomorphology 108, 127–137. Huang, Y.Q., Cheng, Y.P., 2008. Study on sediment concentration variation in Lijiang River basin. Chinese J. Sediment. Res. 58–62 (in Chinese with English abstract).

5. Conclusions and implications The present large-scale study in the riparian zone of the Lijiang River watershed (N = 101) found that the SOCD in the riparian non-bamboodominated woodlands was higher than that in the adjacent grasslands, croplands and orchards by 32.91%, 27.58% and 73.61%, respectively. This indicates that the riparian non-bamboo-dominated woodlands also had a higher SOC storage than the adjacent grasslands and farmlands in the karst region. Deforestation and the conversion of non-bamboo-dominated woodlands to farmlands can lead to a large SOC loss. However, the SOCD in the riparian bamboo-dominated woodland was significantly lower than that in non-bamboo-dominated woodlands, with no significant difference from the grasslands and croplands. This suggests that the planting of bamboo cannot promote SOC sequestration in the riparian zone. We found a close relationship between the SOCD and the soil texture in the riparian non-bamboo-dominated woodlands, where the SOCD showed a significantly positive correlation with the soil clay content, and the soil clay content could account for 36.0% of the SOCD variation. The soil root biomass and plant litter content showed a significantly positive correlation with the SOCD in the riparian nonbamboo-dominated woodland. That is the accumulation of soil root biomass and plant litter could effectively increase the SOC accumulation. Future riparian woodland management should strengthen the management and protection of riparian non-bamboo-dominated woodlands in the karst regions and decrease the SOC loss by the conversion of riparian woodlands to farmlands and deforestation, fulfilling the role of riparian woodlands in sequestrating SOC to mitigate the increasing atmospheric CO2 concentrations. Although bamboo plantation in the riparian zone of the Lijiang River watershed can provide a landscape and economic value, the present management mode of bamboo-dominated woodlands cannot promote the soil of bamboo-dominated woodlands to sequestrate OC and can even cause SOC losses. Thus, the present bamboo planting and management mode needs to be adjusted and new orchard management models that are oriented toward a harmonious ecosystem between the benefits of the landscape and SOC sequestration should be established. Acknowledgments This study was financially funded by Key Projects in the National Science & Technology Pillar Program of China during the Twelfth Five-year Plan Period (2012BAC16B03). The authors declare that they have no conflicts of interest.

Y. Qin et al. / Catena 153 (2017) 21–29 Jose, S.B., 2009. Agroforestry for ecosystem services and environmental benefits: an overview. Agrofor. Syst. 76, 1–10. Larpkern, P., Moe, S.R., Totland, Ø., 2011. Bamboo dominance reduces tree regeneration in a disturbed tropical forest. Oecologia 165, 161–168. Leopold, L.B., Wolman, M.G., Miller, J.P., 1964. Fluvial Processes in Geomorphology. Courier Dover Publications, Mineola, NY. Liu, S., Zhang, W., Wang, K., Pan, F., Yang, S., Shu, S.Y., 2015. Factors controlling accumulation of soil organic carbon along vegetation succession in a typical karst region in Southwest China. Sci. Total Environ. 521, 52–58. Luo, Y., Su, B.O., Currie, W.S., et al., 2004. Progressive nitrogen limitation of ecosystem responses to rising atmospheric carbon dioxide. Bioscience 54, 731–739. Margalef, R., 1958. Information theory in ecology. General Systematics 3, 36–71. Mitsch, W.J., Gosselink, J.G., 2015. Wetlands. fifth ed. Wiley, Hoboken, NJ. Morris, S.J., Bohm, S., Haile-Ma'riam, S., Paul, E.A., 2007. Evaluation of carbon accrual in afforested agricultural soils. Glob. Chang. Biol. 13, 1145–1156. Naiman, R.J., Decamps, H., Pollock, M., 1993. The role of riparian corridors in maintaining regional biodiversity. Ecol. Appl. 3, 209–212. Nelson, D.W., Sommers, L.E., Miller, R.H., Keeney, D.R., 1982. Total carbon, organic carbon, and organic matter. In: Page, A.L. (Ed.), Methods of Soil Analysis, Part 2, Chemical and Microbial Properties. Agronomy Society of America, Agronomy Monograph 9, Madison, Wisconsin, pp. 539–552. Ren, Y., Wang, D.M., Xin, Z.B., 2014. Classification and ecological characteristics analysis of vegetation arrangement type in land-water ecotone of Li River. Acta Ecol. Sin. 34, 4423–4434 (in Chinese with English abstract). Ricker, M.C., Lockaby, B.G., 2015. Soil organic carbon stocks in a large eutrophic floodplain forest of the Southeastern Atlantic coastal plain, USA. Wetlands 35, 291–301. Ricker, M.C., Stolt, M.H., Zavada, M.S., 2014. Comparison of soil organic carbon dynamics in forested riparian wetlands and adjacent uplands. Soil Sci. Soc. Am. J. 78, 1817–1827. Rieger, I., Lang, F., Kowarik, I., Cierjacks, A., 2014. The interplay of sedimentation and carbon accretion in riparian forests. Geomorphology 214, 157–167. Rother, D.C., Rodrigues, R.R., Pizo, M.A., 2009. Effects of bamboo stands on seed rain and seed limitation in a rainforest. For. Ecol. Manag. 257, 885–892. Ruffing, C.M., Dwire, K.A., Daniels, M.D., 2016. Carbon pools in stream-riparian corridors: legacy of disturbance along mountain streams of south-eastern Wyoming. Earth Surf. Process. Landf. 41, 208–223.

29

Schilling, K.E., Palmer, J.A., Bettis III, E.A., Jacobson, P., Schultz, R.C., Isenhart, T.M., 2009. Vertical distribution of total carbon, nitrogen and phosphorus in riparian soils of Walnut Creek, southern Iowa. Catena 77, 266–273. Seth, W., 1999. A Review of the Scientific Literature on Riparian Buffer Width, Extent and Vegetation. Institute of Ecology, University of Georgia, Athens, Georgia. Shannon, C.E., Weaver, W., 1963. The Mathematical Theory of Communication. University of Illinois Press, Urbana, USA. Sutfin, N.A., Wohl, E.E., Dwire, K.A., 2016. Banking carbon: a review of organic carbon storage and physical factors influencing retention in floodplains and riparian ecosystems. Earth Surf. Process. Landf. 41, 38–60. Tockner, K., Stanford, J.A., 2002. Riverine flood plains: present state and future trends. Environ. Conserv. 29, 308–330. Umali, B.P., Oliver, D.P., Forrester, S., Chittleborough, D.J., Huston, J.L., Kookana, R.S., Ostendorf, B., 2012. The effect of terrain and management on the spatial variability of soil properties in an apple orchard. Catena 93, 38–48. Vidon, P., Allan, C., Burns, D., et al., 2010. Hot spots and hot moments in riparian zones: potential for improved water quality management. J. Am. Water Resour. Assoc. 46, 278–298. Vockenhuber, E.A., Scherber, C., Langenbruch, C., Meißner, M., Seidel, D., Tscharntke, T., 2011. Tree diversity and environmental context predict herb species richness and cover in Germany's largest connected deciduous forest. Perspectives in Plant Ecology, Evolution and Systematics. 13, 111–119. Wei, Y.G., 2004. The floristic features and the relationship between the plant community and the landscape on the bank of Li River. Guihaia 24, 508–514 (in Chinese with English abstract). Were, K., Bui, D.T., Dick, Ø.B., Singh, B.R., 2015. A comparative assessment of support vector regression, artificial neural networks, and random forests for predicting and mapping soil organic carbon stocks across an Afromontane landscape. Ecol. Indic. 52, 394–403. Yuan, Y., Bingner, R.L., Locke, M.A., 2009. A review of effectiveness of vegetative buffers on sediment trapping in agricultural areas. Ecohydrology 2, 321–336. Zhou, B.Z., Fu, M.Y., Xie, J.Z., Yang, X.S., Li, Z.C., 2005. Ecological functions of bamboo forest: research and application. J. For. Res. 16, 143–147.