Differential responses of litter decomposition to increased soil ...

1 downloads 0 Views 307KB Size Report
with contrasting life forms and tissue chemistry in a typical steppe of Inner Mongolia, China. ... nutrient dynamics of the semi-arid grasslands of Inner Mongolia.
Applied Soil Ecology 34 (2006) 266–275 www.elsevier.com/locate/apsoil

Differential responses of litter decomposition to increased soil nutrients and water between two contrasting grassland plant species of Inner Mongolia, China Ping Liu a,b, Jianhui Huang a, Xingguo Han a, Osbert J. Sun a,*, Zhiyong Zhou a,b a

Laboratory of Quantitative Vegetation Ecology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China b Graduate University of Chinese Academy of Sciences, Beijing 100049, China Received 1 August 2005; received in revised form 22 December 2005; accepted 28 December 2005

Abstract Soil chemistry and physical conditions are the key factors controlling litter decomposition. We studied the effects of increased soil nitrogen (N), phosphorus (P), and water on the decomposition rates and associated nutrient dynamics of two dominant grassland plant species (i.e. Allium bidentatum Fisch. ex Prokh. & Ikonn.-Gal. and Stipa krylovii Roshev.) with contrasting life forms and tissue chemistry in a typical steppe of Inner Mongolia, China. The treatments included addition of urea at the rates equivalent to 0, 8, 16, and 32 gN/m2, and additions of mixed urea and triple superphosphate at the rates equivalent to 0, 8 gN/4 gP, 16 gN/8 gP, and 32 gN/16 gP/m2, with and without water added. We found marked differences between the two species in the rates, as well as in the responses to water and addition rates of N and P, of litter decomposition. Additions of N alone or in mixture with P stimulated the rate of litter decomposition in both species. Adding water significantly increased the values of decay constant, k, in A. bidentatum, but not in S. krylovii. N and P concentrations in litters of A. bidentatum and S. krylovii all increased corresponding to increases in the rates of N or mixed N and P additions. Our results clearly indicate that the decomposition of high quality litter is more likely to be limited by soil moisture regimes, while that of low quality litter is more sensitive to nutrient availability. Our findings suggest that plant species with different litter qualities should be taken into consideration when we are to model the carbon cycle and nutrient dynamics in grassland ecosystems and that A. bidentatum is expected to contribute more than S. krylovii to the carbon cycle and nutrient dynamics of the semi-arid grasslands of Inner Mongolia. # 2006 Elsevier B.V. All rights reserved. Keywords: Litter decomposition; Grassland ecosystems; Allium bidentatum; Stipa krylovii

1. Introduction Decomposition of plant litter plays an important role in nutrient cycling and carbon (C) fluxes of the terrestrial ecosystems (Swift et al., 1979; Berg and McClaugherty, 1989; Sun et al., 2004). The rate and process of litter decomposition greatly influence soil * Corresponding author. Tel.: +86 10 62836510. E-mail addresses: [email protected], [email protected] (O.J. Sun). 0929-1393/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apsoil.2005.12.009

development and the availability of nitrogen (N), phosphorus (P) and other nutrients to plants and soil microorganisms (Huang et al., 1998; Liu et al., 2000). Understanding the key factors and processes that control the rate of litter decomposition under different environmental conditions and in different habitats is therefore fundamental to quantitative analysis of C and nutrient cycling of terrestrial ecosystems. Decomposition is primarily driven by microbial activities and can be best predicted by environmental factors such as temperature and precipitation, as well as

P. Liu et al. / Applied Soil Ecology 34 (2006) 266–275

litter quality (Meentemeyer, 1978; Berg et al., 1993; Aerts, 1997; Taylor et al., 1989; Wang et al., 2000; Moretto et al., 2001), but soil chemistry and physical conditions can also influence the rate of litter decomposition (Gijsman et al., 1997; Seneviratne et al., 1999). As decomposer microbes require nutrients from either litter material or surrounding soils to maintain their life activities (Ocio et al., 1991; Sinsabaugh et al., 1993), soil nutrient availability has long been suggested as one of the controlling factors affecting the rate of litter decomposition (Swift et al., 1979). However, results of the studies concerning the effects of increased soil N and P on the rate of litter decomposition and nutrient dynamics to date have been controversial. For example, some studies (Hunt et al., 1988; Fenn, 1991; Ostertag and Hobbie, 1999) found that increased N and P could stimulate litter decomposition, while others found no (i.e. Prescott et al., 1999; Dukes and Field, 2000) or depressing effects (So¨derstro¨m et al., 1983; Magill and Aber, 1998). Berg et al. (1982) and Berg and Matzner (1997) found a positive response to N in the initial decomposition phase but a negative response in the later stages. Kwabiah et al. (1999) suggested that responses of plant litter decomposition to soil nutrients were determined by litter quality. Such inconsistency in the relationships between litter decomposition and soil nutrients, therefore, calls for continued investigations of the subject. Water availability can influence the rates of litter decomposition and nutrient release through its effects on the activities of the decomposer communities (Orchard and Cook, 1983; Berg, 1986; Clein and Schimel, 1994). Water supply in the form of rainfall can also affect decomposition by facilitating leaching and breakdown of surface litter (Swift et al., 1979). Moreover, water availability may affect litter decomposition indirectly by altering the litter quality in terms of lignin and nutrient concentrations of plants (Pastor and Post, 1988; Austin and Vitousek, 2000). Global change will inevitably alter nutrient cycling processes in terrestrial ecosystems along with changes in C and water fluxes. The accelerated rates of N mineralization likely to occur with global climate warming (Rustad et al., 2001) and increased N deposition due to anthropogenic activities (Bobbink et al., 1998) could result in N enrichment under many land use and land cover types, including grassland ecosystems. The global precipitation pattern is anticipated to change under the climate change scenarios, especially in arid and semi-arid areas where summer precipitation is expected to increase (Melillo, 1990). All

267

these changes would potentially result in changes in litter decomposition that could affect C and nutrient cycles, and affect how we simulate the C cycling process in the context of global change. However, there is a lack of regionally based, species-specific information on the rates of litter decomposition for most of the terrestrial ecosystems in China. The increased intensity of land use on the grasslands of northern China has caused severe land degradation since the late 1970s, resulting in decreased productivity and ecosystem stability. Much of the area is now in various stages of degradation, where soil N, P and water are considered to be the principal limiting factors to net primary productivity. The recycling of nutrients through plant litter decomposition is an essential mechanism to maintain productivity (Tiessen et al., 1994). To examine how increased soil nutrient availability, predominantly N and P, and water would affect the litter decomposition rates and nutrient dynamics of two dominant grassland plant species with contrasting life forms and tissue chemistry (i.e. Allium bidentatum and Stipa krylovii), we conducted two field experiments with varying addition rates of N, P, and water. Our objectives were to determine: (1) how increased soil N would affect the decomposition rates and nutrient dynamics of A. bidentatum and S. krylovii, and (2) how simultaneous increases in soil N and P, with or without increased water availability, would affect the rates of decomposition of the two species. Our underlying hypotheses were that species with contrasting life forms and tissue chemistry would differ significantly in the rates of litter decomposition, and that increased soil nutrient and water would stimulate litter decomposition in the degraded semi-arid grassland ecosystems. 2. Materials and methods 2.1. Study site This study was conducted at a field site of the Duolun Restoration Ecology Experimentation and Demonstration Station in Inner Mongolia, China (latitude 428020 N; longitude 1168160 E; altitude 1344 m a.s.l.). The mean annual, minimum, and maximum air temperatures for the area are 1.6, 18.3, and 18.7 8C. The mean annual precipitation is 385 mm, occurring mainly from July to September (accounting for 67% of the total), with Penman evaporative potential of 1748 mm. Soils of the region are commonly referred to as chestnut type (Calcic Kastanozems) in the Chinese classification, and are classified as Calcic–orthic Aridisol in the US soil taxonomy classification system (Yuan et al., 2005).

268

P. Liu et al. / Applied Soil Ecology 34 (2006) 266–275

The area has suffered from severe degradation since 1979 as a result of overgrazing. The primary vegetation is of the typical steppe, with perennial herb A. bidentatum and bunch grass species S. krylovii being among the dominant species of the plant community.

the time of initial deployment to determine the ratio between air-dried mass and oven-dried mass. This ratio was used to convert the initial air-dried mass of the litter to oven-dried mass. 2.4. Chemical analysis

2.2. Experimental design and treatments We conducted two decomposition experiments on a fenced site from July and October 2004, both lasting for 100 days. The first experiment contained four levels of N addition rates (0, 8, 16, and 32 gN/m2; designated as N0, N8, N16, and N32, respectively), and the second experiment contained four addition rates of mixed N and P (0, 8 gN/4 gP, 16 gN/8 gP, and 32 gN/16 gP/m2; designated as N0P0, N8P4, N16P8, and N32P16, respectively) with and without increased water availability. Both of the experiments were deployed based on a randomized block design with four blocks and each plot size of 4 m  4 m. N was added as granular urea, and P in the form of granulated triple superphosphate, which were applied 5 days before the litterbags were deployed. In treatments with increased water availability, water was applied on 15 July, 10 August, 29 August, and 22 September, following major rain events in the previous week. For each water application, 15 mm of water was added, simulating an average increase of rainfall by 30% of the corresponding period. 2.3. Litter decomposition In early July 2004, litters of A. bidentatum (predominantly leaves) and S. krylovii (predominantly culms) were collected from the study area, including freshly fallen and senesced tissues attached to the plants. After being air-dried to constant mass, they were clipped into fragments of 10 cm in length, and placed into the 10 cm  15 cm polyethylene litterbags (mesh size 1 mm2) that each contained 3 g of air-dried litter material. Each treatment plot contained nine litterbags of the same species. A total of 864 litterbags were prepared and deployed onto the treatment plots on 9 July 2004. The remaining litter was retrieved 35, 70, and 100 days after initial litterbag deployment. For each sampling time, three litterbags for each species from each treatment plot were collected. In the laboratory, extraneous matter such as other plant materials, rocks and small animals were handpicked from the litterbags. The retrieved litters were then oven-dried at 70 8C for 48 h, to determine the remaining dry mass. Five samples for each litter type were oven-dried at 70 8C for 48 h at

After determination of the dry mass, litters of the same plant species from the three within-plot replications were pooled for chemical analysis. Total C, N and P concentrations were determined for the final samples from the two litter decomposition experiments. Total C was measured by the standard method of wetcombustion, total N by semi-micro Kjeldahl method and total P by molybdenum blue colorimetric method (Bao, 1999). Five litter samples for each plant species were also analyzed for total C, N and P to determine the initial litter chemistry. 2.5. Data analysis The value of decay constant, k, was determined by fitting the following exponential function (Olson, 1963): xt ¼ ekt x0 where, xt is the remaining litter mass after a given time period t, x0 is the initial litter mass. The remaining litter nutrients were calculated by multiplying the sample mass by the respective nutrient concentration. Tests of significance of the remaining mass were performed by analysis of variance (ANOVA; three-way without water treatment and four-way with water treatment). Data analyses were performed using procedures of SPSS (v.11.0). The least significant difference (LSD) was used for comparisons of means with confidence level of P < 0.05. 3. Results 3.1. Initial litter chemistry Total C contents in the initial litters were similar between A. bidentatum and S. krylovii (Table 1). Allium bidentatum had twice the N and P concentrations, and half the C:N and C:P ratios, of S. krylovii (Table 1). 3.2. Species difference in litter decomposition Allium bidentatum litter decomposed faster than S. krylovii in both experiments (Fig. 1). After 100 days of decomposition, the remaining litter mass was 70% of

P. Liu et al. / Applied Soil Ecology 34 (2006) 266–275

269

Table 1 Initial litter chemistry of Allium bidentatum and Stipa krylovii in the degraded typical steppe of Inner Mongolia, China Species

C (g/kg)

N (g/kg)

P (g/kg)

C:N ratio

C:P ratio

Allium bidentatum Stipa krylovii

481.9  6.0 486.3  7.2

5.0  0.3 2.8  0.2

0.42  0.02 0.23  0.01

97  6 174  15

1273  65 2720  101

Values are means  S.E. (n = 5).

the initial in A. bidentatum, but >80% in S. krylovii. The patterns of litter decomposition over the duration of the experiment were well described by the exponential model, xt/x0 = ekt (values of r2 ranged between 0.89 and 0.99, and P < 0.05; Table 2). The values of decay

constant, k, of A. bidentatum were about twice of S. krylovii across all the treatments (Table 2). 3.3. Effects of nutrient addition and water on litter decomposition In the experiment I, addition of N alone to the substrate soils only slightly affected the rate of litter decomposition in both A. bidentatum and S. krylovii as shown by three-way ANOVA and changes in the k values, with two species showing the similar patterns of responses (Tables 2 and 3). Four-way ANOVA showed that the litter mass remaining was significantly (P < 0.05) affected by retrieval time, plant species, water treatment, addition rates of mixed N and P, and some of the interactions in the experiment II (Table 4). Increased rates of N and P Table 2 Decay constant (k; y1, r2 range 0.89–0.99, P < 0.05) for litters of Allium bidentatum and Stipa krylovii as affected by additions of N alone (experiment I) or in mixture with P with or without adding water (experiment II) in the degraded typical steppe of Inner Mongolia, China. Treatment

Fig. 1. Effects of N addition or concurrent additions of N and P with or without water added on litter decomposition of Allium bidentatum (open symbols) and Stipa krylovii (closed symbols). Vertical bars represent standard errors (n = 4). *P < 0.05; **P < 0.01; ***P < 0.001.

k Allium bidentatum

Stipa krylovii

Experiment I N0 N4 N8 N16

1.20  0.07 1.24  0.10 1.30  0.10 1.33  0.08

a a a a

0.53  0.07 0.58  0.07 0.63  0.10 0.67  0.06

b ab ab a

Experiment II No water added N0P0 N8P4 N16P8 N32P16

1.20  0.07 1.23  0.11 1.37  0.05 1.48  0.13

c bc ab a

0.63  0.06 0.69  0.02 0.77  0.07 0.85  0.10

c bc ab a

Water added N0P0 N8P4 N16P8 N32P16

1.37  0.07 1.41  0.15 1.48  0.09 1.56  0.09

b ab ab a

0.63  0.01 0.71  0.02 0.79  0.04 0.87  0.05

d c b a

The values of k are means  S.E. (n = 4), which were derived based on exponential decay model, xt/x0 = ekt. Data in the same column followed by the same letter are not significantly different (P < 0.05).

270

P. Liu et al. / Applied Soil Ecology 34 (2006) 266–275

Table 3 F and P (shown in parentheses) values from three-way ANOVA for mass remaining (M, % of initial mass) and two-way ANOVA for residual N and P (% of initial mass) during decomposition of litter materials in soils with N addition (experiment I), with retrieval time, plant species and N addition (fertilizer) as main effects Source of variation

d.f.

F M

Time (T) Species (S) Fertilizer (F) TS TF SF TSF

2 1 3 2 6 3 6

79.3 286.5 2.4 10.5 0.9 1.4 0.9

(