Introduction
Soil organic matter (SOM) plays a central role in the global biogeochemical
cycles of most major nutrients. The soil C pool is the largest C storage in
the terrestrial ecosystem and the organic C pool consists of more than
90 % of the total soil C pool. Grasslands account for 40.5 % of the
terrestrial area globally, and it is estimated that 34 % of the global
terrestrial organic C was stored in the grasslands (White et al., 2000). In
China, grasslands also cover more than 40 % of the terrestrial surface
and the Inner Mongolia grassland, which is one of the most important animal
husbandry bases, representing more than a quarter of the total grassland area
(National Environment Protection Bureau of China, 2006). Therefore, soil C
and nitrogen (N) cycling in Inner Mongolia grassland has been a hot topic
(Shan et al., 2011; Wang et al., 2014, 2016). Mowing once a year is one of
the common practices in grassland ecosystems, and it is reported that mowing
once a year increases the stocks of soil C and N by facilitating plant
species richness, plant productivity, root biomass and root exudates (Socher
et al., 2012; Cong et al., 2014). However, in order to prepare enough winter
feed for the increased livestock, high-frequency mowing is needed, which
might result in the reduction of plant species diversity and block soil C and
N turnover, as few microbes are able to bear such a degree of disturbance.
Increased plant diversity and enhanced fresh SOC input by mowing once a year
can lead to the degradation of recalcitrant organic compounds by the priming
effect (Fontaine et al., 2011). In addition, different plant species release
diverse organic compounds, and these would have an impact on soil microbial
communities (Dijkstra et al., 2005). It was documented that mowing could
increase the activity of extracellular enzymes to decompose polymeric C
(aromatic polymer from lignin derived from litter or root residue) into
monomers (Steinauer et al., 2015), including simple but resistant C like
alkyl C, a decomposition product which is stable in soil. The stability of
the soil C pool is closely related to the sustainability of soil functions.
However, it has been unclear how stable the SOM is under the different mowing
managements. Therefore, to better assess the ecological significance of
long-term mowing managements, it is necessary to study the impacts of
different mowing managements on the quantity and quality of SOM.
Soil organic matter composition is often used to evaluate the stability of
soil C pools. The fractions of SOM are traditionally classified based on the
assumption of organo-mineral interactions and spatial arrangements of soil
particle size by physical methods (Cao et al., 2011) and relying on their
solubility in acid or base extractants by chemical methods (Olk and
Gregorich, 2006). Generally, labile SOC fractions include water-soluble
organic C (WSOC), microbial biomass C (MBC) and readily
oxidized C (ROC), which are considered to be
early and sensitive indicators of soil quality because they could rapidly
respond to soil management practice (Chen et al., 2017), while humus is
recalcitrant SOM. These fractions are extracted by different extractants.
Spectroscopy is a powerful tool for identifying the chemical structures of
SOM as soil samples are measured directly rather than determined after a
series of extractions which might alter the nature of SOM. Fourier-transform
infrared spectroscopy (FTIR) and nuclear magnetic resonance (NMR) are widely
used to study the chemical composition of organic matter (Olk, 2006; Mao et
al., 2008; Zhou et al., 2014).
Advanced solid-state NMR, i.e. cross-polarization magic angle spinning
13C-NMR (CPMAS 13C-NMR) applied in characterizing chemical
structures of SOM, is also an important approach to reveal the essential
changes in SOM formation and degradation in paddy fields and forest
ecosystems (Zhou et al., 2014; Panettieri et al., 2014; Zhang et al., 2015)
and in litter and wood decomposition processes in these ecosystems (Sanaullah
et al., 2012; Bonanomi et al., 2013; Hu et al., 2017). This approach can
provide the information of SOM structure noninvasively without using
solvents. Generally, alkyl C (45–0 ppm), N-akyl/methoxyl C (60–45 ppm),
O-alkyl C (90–60 ppm), di-O-alkyl C (110–90 ppm), aryl C
(140–110 ppm), phenolic C (165–140 ppm) and carboxyl and carbonyl C
(210–165 ppm) are identified in detail, from the spectra of 13C-NMR of
soil samples (Baumann et al., 2009, 2013; Zhao et al., 2012). This functional
C can also be generally grouped into four groups: carbonyl C (210–165 ppm),
aromatics C (165–110 ppm), substituted alkyl C (110–45 ppm), and alkyl C
(45–0 ppm) (Plaza et al., 2013; Zhao et al., 2012; Boeni et al., 2014). In
addition, alkyl, N-akyl/methoxyl, aryl and phenolic C are often derived
from lignin, while O-alkyl and di-O-alkyl C are generally included in
polysaccharide (carbohydrates) (Preston et al., 1998; Bonanomi et al., 2013). By
analysing the functional C
composition, the nature of SOM can be better understood and the quality and
quantity of SOM can be more exactly evaluated. However, the cost of
solid-state 13C-NMR is high, especially for complex soil samples,
because of the length of time it takes to identify the chemical structure.
To better understand and evaluate the quality of SOM, elemental analysis
(EA) and FTIR are often combined to help get more accurate information (Mao
et al., 2008; Zhou et al., 2014). In grassland ecosystems, these tools are
also used to study SOM stocks and quality (Baumann et al., 2016; Knicker et
al., 2012). However, there are no reports of the effects of mowing and
mowing frequency on the chemical structure of the whole SOM, which reflects
the nature of SOM. In this study, we combined advanced solid-state NMR with
traditional methods to investigate the quality and quantity of the grassland
soil organic C under different mowing managements. The objective of this
study was to investigate the impacts of long-term mowing practices on the
chemical composition of SOM and evaluate the stability of the grassland soil
carbon pools under different mowing frequencies.
Materials and methods
Site description and experimental design
The study site was located in the Xilingol region of Inner Mongolia
(43∘269′–44∘089′ N and
116∘049′–117∘059′ E) in northern China. It has a
temperate semiarid climate, with an annual mean temperature of
0.5 ∘C and annual average precipitation of 350 mm, most of which
falls during the summer. The annual potential evapotranspiration ranged from
1600 to 1800 mm. The soil was Calcic-orthic Aridisol according to the US
soil taxonomy (or sandy-loam dark chestnut soil in the Chinese classification
system) (Baoyin et al., 2014) and, in the profile, there was a humus layer of
20–30 cm and a calcic layer at ca. 50 cm depth (Jiang, 1988). The
characteristic vegetation of this region was Leymus chinensis
(L. chinensis), accounting for 55 ± 15 %
(mean ± standard deviation) of total herbage yield. Other species in
order of decreasing proportion of total herbage yield are tall bunchgrasses
(mostly Stipa grandis and Agropyron michnoi), short
bunchgrasses (Cleistogenes squarrosa and Koeleria cristata
(L.) Schrad) and sedge (Carex korshingski Kom.), forbs and legumes
(Baoyin et al., 2014). The growing season usually started in May and ended in
September.
The long-term mowing experiment has been carried out since 2001 in a
permanent enclosure by the Inner Mongolia Grassland Ecosystem Research
Station of the Chinese Academy of Sciences. The enclosure for the mowing
experiment was divided into 12 plots (24 m × 20 m for each plot).
There were four treatments, each with three replicates. The four treatments
were unmown (M0), mowing once every two years (M1/2) in August, mowing once a
year (M1) in August when aboveground biomass of L. chinensis reached
the peak, and mowing twice every year (M2) in June when the palatability of
L. chinensis was best for the livestock and in September when
L. chinensis was withered.
Soil sample collection
Soil samples were collected from 0 to 10 cm depth using a soil auger (7 cm
in diameter and 10 cm in depth) in October 2013 at the end of the growing
season and all plots experienced the grass cutting in this year. Five soil
cores were collected from each plot at random locations and they were
combined and mixed thoroughly to form a composite sample. Visible roots and
litter residues and large soil fauna in the soil samples were removed. The
soil samples of around 1 kg were put into ziplock bags and transported to
the lab on ice quickly. In the lab, the soil samples were passed through a
2 mm sieve and subsampled into two parts. One part was air-dried for basic
physical and chemical properties analysis; the other fresh part was used to
analyse the fractions and chemical composition of SOM as soon as possible
and, if this part could not be analysed in a short time, they were stored at
-20 ∘C to avoid the impacts of storage temperature on the
indicators determined, especially the microbial indicators.
Measurements of bulk soil basic properties
Soil pH was measured using a water-to-soil ratio of 2.5 : 1. Soil moisture
content was determined by oven drying for 16 h to a constant mass at
105 ∘C. The content of soil organic C (SOC) and total nitrogen (TN),
alkali-hydrolysable N (AN), Olsen phosphorus (Olsen P) and net N
mineralization was determined, referring to Kalembasa and Jenkinson (1973),
Bremner (1965), Bao (2000)
and Lin (2010), respectively.
Soil organic matter fractionation
Soil microbial biomass carbon (MBC) was extracted using the chloroform
fumigation extraction (Vance et al., 1987; Wu et al., 1990) and determined
using a TOC analyser (Elementar Liqui TOC, Elementar Co., Hanau, Germany).
Water soluble organic carbon (WSOC) was determined using a modified method
(Li et al., 2013). Briefly, WSOC was extracted from 5.0 g of fresh soil
using a soil-to-water ratio of 1 : 10 at 25 ∘C, and shaken for
30 min at a speed of 250 rpm. The samples were subsequently centrifuged
(1146 × g, 20 min), and then the
supernatant was filtered using a 0.45 µm membrane filter. The
filtrate was measured by the same TOC analyser mentioned above. Soil readily
oxidizable carbon (ROC) was determined and calculated following the detailed
procedure described by Li et al. (2013). The mobile humic acid (MHA) and
calcium humic acid (CaHA) were extracted following the procedure by Mao et
al. (2008). Thirty grams of air-dry soil was used to extract the two humic
fractions and the extracted humic fractions were freeze-dried using a
freeze-drying machine (FD-1C-50, Beijing, China), and then weighted,
respectively.
Calculation formulas of different 13C-NMR indices.
Index
Formula
Reference
A / OA
alkyl C (45–0 ppm) / O-alkyl C (110–60 ppm)
Mathers et al. (2007)
CC / MC
carbohydrate C (90–60 ppm) / methoxyl C (60–45 ppm)
Zhao et al. (2012)
HB / HI
hydrophobic C (45–0 ppm + 165–110 ppm) / hydrophilic C (110–60 ppm + 210–165 ppm)
Spaccini et al. (2002)
Aliphaticity, %
(alkyl C + Substituted C) ⋅ 100 / (alkyl C + substituted C + aromatic C)
Zhao et al. (2012)
Aromaticity, %
aromatic C ⋅ 100 / (alkyl C + substituted C + aromatic C)
Al / Ar
aliphaticity / aromaticity
Lignin C
phenolic C ⋅ 4.5 + methoxyl C
Preston et al. (1998)
Polysaccharide C
1.2 ⋅ (O-alkyl C - phenolic C ⋅ 1.5)
L / P
lignin C / polysaccharide C
Analysis of the chemical composition of soil organic matter
(SOM)
To remove paramagnetic materials (Fe3+, Mn2+) and increase the
signal-to-noise ratio, the soil samples were pretreated with HF (10 %,
v/v) using the procedure detailed in Li et al. (2010), and finally, the SOM
samples were freeze-dried. It is reported that the chemical composition of
SOM was not altered as the C / N was similar before and after the HF
processing (Mao et al., 2008; Zhou et al., 2014), as was the case in this
study (Table S2 in the Supplement), and the C and N content in 10 %
HF-treated SOM samples and bulk soil samples was measured using a CHNS
Elemental Analyzer (Carlo Erba model EA1108, Italy Vario).
Elemental analysis
The elemental composition of 10 % HF-treated SOM samples was determined
using the same CHNS Elemental Analyzer mentioned above. The content of O was
estimated as the ash-free mass less C, H, and N. Ash content was determined
by combustion overnight in a muffle furnace at 500 ∘C (Ussiri and
Johnson, 2003).
FTIR analysis
The FTIR analysis of the SOM samples was conducted on an Avatar 370 FTIR
spectrometer (Thermo Nicolet, USA).
Each sample was prepared by grinding a 2 mg freezing-drying SOM sample with
200 mg oven-dried KBr in a vibrating puck mill, and then mixtures of about
150 mg were compressed into a translucent pellet using a hydraulic
compressor. The pellet was immediately placed on the sample holder, and all
spectra ranging from 4000 to 400 cm-1 were recorded under the
conditions of 4 cm-1 wave number resolution, 25 scans, and pure KBr
spectra as background (Zhou et al., 2014). Absorption peaks or bands were
assigned to organic functional groups following Zhou et al. (2014). Only
peaks or bands in the functional group region from 4000 to 1000 cm-1 of
FTIR spectra were assigned because peaks in the fingerprint region below
1000 cm-1 were difficult to assign and were very complex, usually
overlapping with signals of inorganic soil minerals.
Solid-state CPMAS 13C-NMR analysis
A solid-state 13C-NMR experiment was performed on a Bruker
Avance II 300 (Bruker Instrumental Inc) equipped with a 7 mm CPMAS
(cross-polarization magic-angle-spinning) detector. NMR spectra were acquired
under the conditions of a spectrometer frequency of 75 MHz, a MAS spinning
frequency of 5000 Hz, a recycle time of 2.5 s and a contact time of 2 ms.
The external standard used for chemical shift determination was
hexam-ethylbenzene (methyl at 17.33 ppm). The quantified contribution of
each type of C to the total signal intensity and promotion in CPMAS
13C-NMR spectrum was automatically integrated after the separation of
hexam-ethylbenzene to calculate the area of the peaks which appeared in the
corresponding chemical region using MestreNova software 8.1.0 (Mestrelab,
Research Inc). The 13C-NMR spectrum was assigned into seven regions as
the previous studies (Baumann et al., 2009, 2013; Zhao et al., 2012) and they
were grouped into four main chemical environments according to the 13C
nucleus: carbonyl C (210–165 ppm), aromatics C (165–110 ppm), substituted
alkyl C (110–45 ppm), alkyl C (45–0 ppm) (Plaza et al., 2013; Zhao et
al., 2012; Boeni et al., 2014). The seven assignments of 13C-NMR
spectrum and potential sources of functional groups in each assignment were
showed in Table S1. To better evaluate the quality of C pools, some indices
were calculated following the formula in Table 1.
Basic description of soil properties under different mowing treatments.
Treatment
pH
SOC
TN
Olsen P
AN
Net N mineralization
H2O
g kg-1
mg kg-1
mg N g-1
M0
7.3 ± 0.1 a
17.9 ± 0.6 b
1.5 ± 0.0 ab
1.0 ± 0.15 a
75 ± 0.59 a
194 ± 3.76 b
M1/2
7.3 ± 0.0 a
20.2 ± 1.6 a
1.5 ± 0.4 ab
1.3 ± 0.06 a
86 ± 1.42 a
176 ± 7.51 b
M1
7.3 ± 0.1 a
21.7 ± 0.3 a
1.7 ± 0.0 a
1.2 ± 0.03 a
86 ± 0.00 a
225 ± 2.51 a
M2
7.2 ± 0.0 a
17.8 ± 0.8 b
1.3 ± 0.0 b
1.2 ± 0.03 a
57 ± 0.00 b
127 ± 7.50 c
M0, unmown; M1/2, mowing once every second year; M1, mowing once
a year; M2, mowing twice a year. The value was the mean ± S.E., n=3. SOC, soil organic carbon; TN, total nitrogen; Olsen P, Olsen phosphorus;
AN, alkali-hydrolysable nitrogen; net N mineralization, net nitrogen
mineralization. Different lowercase letters in the same column indicated that
the difference between treatments reaches a 5 % significance level.
Data analysis
Data were statistically analysed using SPSS 21.0 by one-way analysis of
variance (ANOVA), and means were separated by Duncan's multiple range test at
5 % level. The figures were created using Origin 8.1 and the data were
the mean values (n=3). Linear regression analysis was conducted after the
Pearson product–moment correlation analysis by a two-tailed test in SPSS
21.0 using the data in all mowing treatments (n=9) except for the unmown
(M0).
Effect of different mowing managements on bulk SOM fractions.
Treatment
MBC
WSOC
ROC
MHA
CaHA
HA / SOM
mg kg-1
g kg-1
%
M0
139.0 ± 9.81 b
98.6 ± 9.42 a
7.3 ± 0.65 a
6.0 ± 0.76 a
14.0 ± 0.87 b
64.8 a
M1/2
167.9 ± 3.70 a
42.4 ± 3.51 b
3.1 ± 0.17 bc
4.6 ± 0.76 a
20.5 ± 0.53 a
72.2 a
M1
144.6 ± 8.09 b
45.6 ± 2.37 b
3.5 ± 0.20 b
6.0 ± 0.55 a
21.5 ± 0.46 a
73.0 a
M2
101.3 ± 6.23 c
38.8 ± 5.51 b
2.3 ± 0.12 c
3.9 ± 0.57 b
12.9 ± 0.89 b
53.1 b
M0, unmown; M1/2, mowing once every second year; M1, mowing once
a year; M2, mowing twice a year. The value was the mean ± S.E., n=3. WSOC, water soluble organic carbon. MBC, microbial biomass carbon. ROC,
readily oxidized carbon. MHA, mobile humic
acid. CaHA, calcium humic acid. SOM, soil total organic matter.
HA = MHA + CaHA.
Results
Basic properties of bulk soil and net N mineralization
Soil pH was around 7.3 and was not affected by long-term mowing (Table 2).
However, long-term mowing had a significant impact on soil nutrient
concentrations. Compared with M0 (unmown), mowing once every second year
(M1/2) and mowing once a year (M1) significantly increased SOC content (P<0.05) while the SOC content in M2 was similar to that in M0 (P>0.05). The TN content in M1 was the highest and significantly
higher than that in treatment M2. The total N content in M2 was also
significantly lower than those in the other two treatments (M1/2 and M0).
Soil Olsen P contents in all the treatments were very low, around
1.2 mg kg-1, and no significant difference was observed between the
treatments (P>0.05). The AN content in the soil in M2 was significantly
lower than those in the other treatments (P<0.05), while there was no
significant difference between the other treatments (P>0.05). Net N
mineralization in M1 was significantly greater than that in the other
treatments, and it was significantly lower in M2 than that in other
treatments (P<0.05).
Elemental composition of SOM from surface soils in grassland soil
with different mowing frequencies.
Treatment
Elemental composition, %
Atom ratios
C
H
N
O
H / C
O / C
M0
3.94 ± 0.03 b
0.63 ± 0.01 a
0.38 ± 0.01 b
0.35 ± 0.02 a
0.16 a
0.09 a
M1/2
3.95 ± 0.04 b
0.51 ± 0.02 b
0.39 ± 0.01 b
0.28 ± 0.02 ab
0.13 b
0.07 b
M1
4.32 ± 0.05 a
0.56 ± 0.02 b
0.41 ± 0.02 a
0.26 ± 0.01 ab
0.13 b
0.06 b
M2
3.28 ± 0.03 c
0.49 ± 0.01 b
0.27 ± 0.02 c
0.25 ± 0.02 b
0.15 a
0.08 a
M0, unmown; M1/2, mowing once every second year; M1, mowing once a year; M2,
mowing twice a year. The value was the mean ± S.E., n=3.
Soil organic matter fractions
Long-term mowing had major impacts on labile C and recalcitrant SOM
(Table 3). Compared with M0, M1/2 significantly increased soil MBC content,
while M2 significantly decreased soil MBC content (P<0.05). WSOC and ROC
contents in all mowing treatments were significantly 50 % lower than
those in M0 (P<0.05). Among different mowing treatments, no difference was
observed in soil WSOC content (P<0.05), while the soil ROC content in M2
treatment was significantly lower than that in M1 (P<0.05). The total
content of both humic fractions (MHA and CaHA) accounted for a major
proportion of SOM, especially in M1, where it reached 73.0 % (Table 3),
and this was significantly higher than that (53.1 %) in M2 (P<0.05).
The CaHA content was about 2–4 times that of MHA across all treatments.
Compared with M0, M2 significantly decreased MHA content (P<0.05) but did
not affect CaHA content significantly (P>0.05). However, M1 and M1/2
significantly increased CaHA content (P<0.05) but did not significantly
affect MHA content. Thus, both MHA and CaHA contents in soils of M1/2 and M1
were significantly higher than that in M2 (P<0.05).
FTIR spectra of bulk SOM under different long-term mowing
managements.
CPMAS 13C-NMR spectra of 10 % HF pretreated SOM.
(a) CPMAS 13C-NMR spectra of 10 % HF pretreated SOM under
different long-term mowing managements. (b) Detailed C functional
groups in different chemical shifts.
Chemical structure of SOM
Parameters of the elemental composition of the SOM were shown in Table 4. The
content of hydrogen (H) and oxygen (O) varied from 0.49 to 0.63 % and
from 0.25 to 0.35 %, respectively. Compared with M0, all mowing
treatments significantly decreased the H content, and M2 also significantly
decreased O content (P<0.05). The ratios of H / C and O / C
varied from 0.13 to 0.16 % and from 0.06 to 0.09 %, respectively, and
the H / C and O / C ratios in M1/2 and M1 were significantly lower
than M2 or M0 (P<0.05).
The FTIR spectra of the SOM extracted from the grassland soil under different
mowing treatments were shown in Fig. 1. The spectra were dominated by the
broad peak around 3406 cm-1, sharp peaks around 1030 cm-1 and
medium sharp peaks around 1653 cm-1, which were ascribed to O–H
stretching in alcohols, carboxylic acids and phenols, C–OH stretching in
carbohydrates, and C=C stretching in aromatics, respectively. The intensity
of other peaks in the FTIR spectra was relatively low. Small peaks at 2928
and closing to 1500 cm-1 due to aliphatic C–H stretching in
CH2 / CH3 and amide N–C / amino–NH vibrations, and
aliphatic C–H bending in CH2 / CH3, respectively, were found
in all treatments. However, only some small differences in the intensity of
the peaks shown in the FTIR spectra
were shown qualitatively between different treatments. The intensity of the
peak at 2928 cm-1 in M1 was stronger than that in M0 and M2, while the
intensity of the peak at 1030 cm-1 in M1 was weaker than that in M0 and
M2 treatments (Fig. 1 and Table 3).
Percentages of total special spectral areas of different functional
groups obtained by quantitative CPMAS 13C-NMR for soil samples from
grassland soil with different mowing frequencies (%).
Alkyl C
Substituted alkyl C
Aromatics
Carbonyls
Treatment
45–0 ppm
60–45 ppm
90–60 ppm
110–90 ppm
140–110 ppm
165–140 ppm
210–165 ppm
Alkyl
N-alkyl/methoxyl
O-alkyl
di-O-alkyl
Aryl
O-aryl
Carboxyl and carbonyl
M0
24.6 ± 0.12 c
11.2 ± 0.06 c
27.0 ± 0.17 a
9.4 ± 0.06 a
11.7 ± 0.08 b
4.6 ± 0.04 c
11.1 ± 0.17 b
M1/2
27.6 ± 0.20 a
12.9 ± 0.05 b
22.5 ± 0.06 b
8.6 ± 0.09 b
13.4 ± 0.09 a
5.7 ± 0.06 b
9.3 ± 0.06 c
M1
27.9 ± 0.23 a
13.4 ± 0.06 a
22.3 ± 0.15 b
8.1 ± 0.06 c
13.6 ± 0.06 a
5.6 ± 0.06 b
9.1 ± 0.06 c
M2
25.4 ± 0.12 b
12.3 ± 0.25 bc
21.7 ± 0.09 c
7.5 ± 0.07 c
11.8 ± 0.12 b
6.6 ± 0.07 a
14.7 ± 0.15 a
M0, unmown; M1/2, mowing once every second year; M1, mowing once a year; M2,
mowing twice a year. The value was the mean ± S.E., n=3.
CPMAS 13C-NMR indices of SOM from surface soils in grassland soils with
different mowing frequencies.
Treatment
Lingin-C
Polysaccharide-C
L / P
Aliphaticity
Aromaticity
Al / Ar
A / OA
HB / HI
CC / MC
% of SOC
%
M0
36.4 c
13.9 b
2.61 b
79.3 a
20.7 c
3.84 a
0.91 b
0.76 b
2.23 a
M1/2
38.6 b
16.7 a
2.30 c
78.9 b
21.1 b
3.74 b
1.23 a
0.88 a
1.74 b
M1
38.2 b
16.9 a
2.26 c
78.5 c
21.5 a
3.65 c
1.25 a
0.87 a
1.66 c
M2
42.0 a
14.2 b
2.97 a
79.4 a
20.6 c
3.85 a
1.17 b
0.81 ab
1.76 b
M0, unmown; M1/2, mowing once every second year; M1, mowing once
a year; M2, mowing twice a year. L / P, lignin / polysaccharide.
A / OA, alkyl C / O-alkyl C. HB / HI, hydrophobic
C / hydrophilic C. CC / MC, carbohydrate C / methoxyl C.
Al / Ar, aliphaticity / aromaticity.
Figure 2 showed the 13C-NMR spectra of the SOM extracted from the
grassland soil with different mowing managements (Fig. 2a) and the detailed C
functional groups represented by the peaks in the 13C-NMR spectra
(Fig. 2b) were shown. In all spectra, the alkyl C (45–0 ppm) and
substituted alkyl C (110–45 ppm) peaks were dominant components in SOC
composition across all the treatments, accounting for 24.6–27.9 and
41.5–47.6 % of the total spectral fractions, respectively (Table 5 and
Fig. 2), followed by aromatic C (165–110 ppm) and carbonyl C
(210–165 ppm) peaks, accounting for 16.3–19.1 and 9.3–13.7 % of the
total spectral fractions, respectively. In the substituted alkyl C, O-alkyl
C (90–60 ppm) was the main fraction, making up more than 50 % of the
substituted alkyl C, while di-O-alkyl (110–90 ppm) only accounted for
less than 21 % of the substituted alkyl C, and N-alkyl/methoxy C was
medium. Compared to M0, mowing significantly increased alkyl C but
significantly decreased substituted alkyl C (except for N-alkyl/methoxyl C,
P<0.05) mainly existing in carbohydrates (Table S1). The proportion of
aromatic C (aryl and O-aryl C, 165–110 ppm) in M1/2 and M1 was
significantly higher than that in M0, while the proportion of carbonyl C
(210–165 ppm) in these two treatments was significantly lower than that in
M0 (P<0.05). Among mowing treatments, alkyl C, substituted C and aryl C in
M2 were significantly lower than those in M1/2 and M1, while O-aryl C and
carbonyl C in M2 were significantly higher than those in M1/2 and M1(P<0.05). The O-alkyl C in M2 was the lowest among all treatments, which was
also consistent with the results of FTIR.
Linear correlation coefficients for relationships among different SOM
fractions and net N mineralization.
SOC
WSOC
MBC
ROC
MHA
CaHA
Net N
mineralization
SOC
1
WSOC
0.11
1
MBC
0.45
0.37
1
ROC
0.36
0.90
0.55
1
MHA
0.48
0.43
0.45
0.92
1
CaHA
0.89
0.41
0.34
0.81
0.82
1
Net N mineralization
0.54
0.08
0.60
0.91
0.83
0.75
1
n=9. The bold denotes the difference was significant at the
level of P<0.05. SOC: soil total organic carbon. The others were the same
as Table 4.
Soil 13C-NMR indices reflecting soil quality directly or indirectly were
calculated and the results were shown in Table 6. Soil lignin C, the
A / OA ratio and the HB / HI ratio in M0 were significantly lower
than those in the mowing treatments (P<0.05), while the CC / MC ratio
in M0 was significantly higher than that in mowing treatments (P<0.05).
Aliphaticity in M1/2 and M1 was significantly lower than that in M0 and M2,
while aromaticity was just the opposite (P<0.05), which resulted in an
Al / Ar ratio in M1/2 and M1 that was significantly lower than that in
M0. There was no difference in aliphaticity, aromaticity and Al / Ar
ratios between M0 and M2 (P>0.05). Among different mowing treatments, most of the 3C-NMR indices in
M1/2 are similar to M1, except that aromaticity in M1/2 was significantly
lower than that in M1, while aliphaticity and the CC / MC ratio in M1/2
were significantly higher than those in M1 (P<0.05). In all of the
indices, lignin C, the L / P ratio and aliphaticity in M2 were
significantly higher than those in both M1/2 and M1, and the other indices
were the opposite (P<0.05).
Summary of the linear correlation for relationships between net N
mineralization, MBC and all the C functional groups of SOC determined by
CPMAS 13C-NMR.
Chemical shifts, region (ppm)
Net N mineralization
MBC
r
P
r
P
Detail assignments
Alkyl C (45–0)
0.46
0.047
0.59
0.039
N-alkyl/methoxyl C (60–45)
0.28
0.615
0.29
0.891
O-alkyl C (90–60)
0.37
0.429
0.27
0.992
di-O-alkyl C (110–90)
0.47
0.326
0.59
0.027
Aryl C (140–110)
-0.24
0.798
0.73
0.011
O-aryl C (165–140)
-0.94
< 0.001
-0.84
< 0.001
Carbonyl C (210–165)
-0.79
0.005
-0.96
0.003
Integrated regions
Unsubstituted alkyl C (45–0)
–
–
–
–
Substituted alkyl C (110–45)
0.70
0.010
0.68
0.014
Aromatics (165–110)
-0.81
0.008
-0.39
0.042
Carbonyls (210–165)
–
–
–
–
n=9. Substituted alkyl C was integrated
into N-alkyl/methoxyl C, O-alkyl C and di-O-alkyl C. Aromatics
integrated aryl C and O-aryl C. In the integrated regions, unsubstituted
alkyl C and carbonyls were the same as alkyl C and carbonyl C in detailed
assignments, respectively.
Variations of SOM fraction and the C functional group in relation
to SOM mineralization and microbial characterization
Soil organic matter content was significantly and positively correlated with
MBC, MHA, CaHA and net N mineralization with r=0.45, 0.48, 0.89, and 0.54
(P<0.05), but was not correlated with WSOC and ROC (P>0.05) (Table 7).
ROC was significantly correlated with WSOC, MBC, MHA, CaHA and net N
mineralization (r=0.55–0.92, P<0.05) and MHA was significantly
correlated with CaHA (r=0.82, P<0.05). Moreover, positive
correlations were found between net N mineralization and MBC, MHA, and CaHA,
with r=0.60, 0.83, and 0.75, respectively (P<0.05).
The relationships between net N mineralization or MBC and the C functional
groups of SOC were shown in Table 8. The results showed that N mineralization
was related to the chemical structure of SOC and to microbial biomass. Net N
mineralization was not significantly related to four detailed CPMAS
13C-NMR regions (N-alkyl/methoxyl C, O-alkyl C, di-O-alkyl C and
aryl C), with r=0.28, 0.37, 0.47 and -0.24, respectively (P>0.05), but was negatively correlated with
O-aryl C (r=-0.94, P<0.001) and carbonyl C (r=-0.79, P<0.01) and the integrated aromatics including aryl C and O-aryl C (r=-0.81, P<0.01). Consistent with net N mineralization, significant
negative correlations were also found between MBC and O-aryl C (r=-0.84, P<0.001), carbonyl C (r=-0.96, P<0.01) and the integrated
aromatics (r=-0.39, P<0.05). However, both net N mineralization and
MBC were positively correlated with alkyl C, with r=0.46 and 0.59,
respectively (P<0.05). Different from net N mineralization, MBC was also
significantly correlated with di-O-alkyl C and aryl C, with r=0.59 and
0.73 (P<0.05), respectively, but was not correlated with
N-alkyl/methoxyl C and O-alkyl C.
Discussion
SOM accumulation impacted by different mowing practices for the long
term
Our results showed that 12-year M1/2 and M1 significantly enhanced SOM
accumulation and increased the soil TN content (Table 2), which agreed with
previous studies (Cong et al., 2014). Mowing (M1) enhanced plant species by
increasing the subordinate plants (Mariotte et al., 2015; Socher et al.,
2012). Enhanced plant species richness promoted plant productivity and
photosynthesis, and thus increased soil carbon and nitrogen stocks in
grasslands by more input of organic C and N derived from more root biomass,
root exudates and N retention and photosynthetic products (Cong et al., 2014;
Gao et al., 2008), which further had a positive feedback to plant
productivity, including legume. Legume was common in grassland and moderate
mowing would stimulate its productivity to increase atmospheric N fixation
(Cardinale et al., 2012), and
N enrichment benefited C accumulation, in turn (Riggs and Hobbie, 2016). In
addition, the significant increase in CaHA content in M1/2 and M1 was the
main and direct reason for SOM accumulation, as the CaHA was the dominant
fraction of SOM (Table 3), which indicated that M1/2 and M1 enhanced the
humus formation. Moderate mowing increased the fungal community abundance and
diversity (Li et al., 2017), and it was reported that fungi could make the
molecular structure of humus more complex (Li, 2012).
Compared to moderate mowing, long-term excessive mowing practice resulted in
herbage productivity decline due to high nutrient removal from the soil and
plant species reduction (Baoyin et al., 2014), which would result in the
decrease in labile SOM fractions (WSOC, MBC and ROC) and relatively labile C
(MHA) contents in M2. Microbes were sensitive to perturbation and thus MBC
was regarded as a reliable indicator of the change in SOC pools caused by
management practices (Fang et al., 2009). The significant reduction of MBC
content was the key biotic reason for soil net N mineralization reduction
(Table 4). Therefore, long-term M2 treatment hampered the soil nutrient
cycling and balance, and should be avoided.
Stability of SOM impacted by different mowing treatments
Different mowing treatments had diverse impacts on the chemical structure of
SOM. The composition of SOM chemical structure directly reflected the
stability of SOM and thus informed the degradability of SOM. The elemental
analysis suggested that long-term mowing practice had major impacts on the
elemental composition of SOM. The lower H / C ratio indicated more
aromatic compounds or higher aromaticity and saturability, and the higher
O / C ratio indicated more carboxyl groups, phenol or carbohydrates with
oxygen (Ma et al., 2001; Steelink, 1985; Kim et al., 1991). 13C NMR
apparently differentiated the lignin C (including alkyl, N-alkyl and aryl
C) and carbohydrate C (including O-alkyl, di-O-alkyl, carbonyl and
carboxyl C) (Hu et al., 2017). Therefore, both elemental analysis and the
quantified analysis of 13C NMR spectra showed that M1/2 and M1 led to a
significant loss of the carbohydrates and accumulation of lignin by more
litter input, which indicated that M1/2 and M1 benefited the stability of
SOM. Previous studies also reported that aryl C at 140–110 ppm was rich in
condensed aromatics and was quite stable in the soil, and its content could
reflect the stability of C pools (Zhou et al., 2014). According to this
conclusion, M2 had little influence on the stability of SOM. However, the
highest content of carbonyl and carboxyl C in M2 suggested that SOM in M2 was
not stable, as compounds that included carbonyl and carboxyl C could be
degraded relatively easily.
The accumulation of lignin and the increase in microbial biomass were the
favourable conditions of humification.
Therefore, the CaHA fraction in SOM increased by 46.9–52.5 % after
12-year M1/2 and M1. This suggested that long-term moderate mowing managements enhanced
the degree of humification of SOM. By contrast, the reduction of litter input
and the significantly decreased microbial biomass in M2 led to the reduction
of humus. Zech et al. (1997) also documented that excessive
human activity resulted in the humic horizons'
disappearance in many tropical regions. This suggested that M2 hindered SOM
humification and disturbed the SOM balance, which might be because the plant
diversity and productivity were limited (Socher et al., 2012; Mariotte et
al., 2013), resulting in lower labile carbon content and less
soil microbial function (Steinauer et al., 2015).
In the CPMAS 13C-NMR indices (Table 7), the A / OA (alkyl
C / O-alkyl C) ratio is generally taken as
a sensitivity index for characterizing the decomposing extent of SOM (Baldock et al., 1997). When the value of the A / OA ratio is
relatively high, it indicates that the degree of decomposition of SOM is
high. In general, alkyl C and O-alkyl C keep a trade-off relationship (Li
et al., 2013). The higher A / OA ratio in M1/2 and M1 could be because
SOM in M1/2 and M1 was difficult to decompose further (Zhao et al., 2012).
Therefore, moderate mowing (M1/2 and M1) enhanced the accumulation of stable
fractions of SOM and recalcitrant chemical structures of SOC, and primed the
degradation of labile C, which suggested that moderate mowing benefited the C
stable sequestration in the semiarid grassland, which was significant to the
grassland C pool. This foundation was reported for the first time in the
grassland ecosystem. The carbohydrate C / methoxyl C (CC / MC) ratio
is a new indicator to reflect the degree of degradation of SOM (Mathers et
al., 2007), and both CC / MC and A / OA ratios showed that the
degradation degree of SOM in M1 was the maximum. In addition to the highest
CaHA content and highest herbage productivity in M1, M1 was the superior
mowing management practice. The aliphaticity / aromaticity (Al / Ar)
ratio is a predictor to reveal the complexity of the chemical composition of
SOM, and the higher the value, the simpler the chemical composition of SOC.
The hydrophobic C / hydrophilic C (HB / HI) ratio was used as a
measure of C chemical recalcitrance, and the higher this value, the more
stable the SOM (Boeni et al., 2014). The increased HB / HI ratios in M1/2
and M1 treatments showed that SOM was more recalcitrant to being mineralized.
Meanwhile, the Al / Ar ratio revealed that M1/2 and M1 increased the
chemical composition complexity of SOM, while M2 had no effect on either the
chemical recalcitrance or complexity of SOM. These further proved that M1/2
and M1 improved the stability of SOM. The higher alkyl C in M1/2 and M1 is
closely associated with the increase in recalcitrant compounds (waxes, resin,
cutin, suberin, peptide side-chain, long-chain aliphatics) (Table S1), mainly
derived from the increased plant materials (Socher et al., 2012; Mariotte et
al., 2013), accompanied by the loss of labile C such as carbohydrates and
polysaccharides, and by the increase in lignin and cellular residues of
microbes (Table 4). It is interesting that lignin C in treatment M2 was
significantly higher than that in other treatments (Table 6), which might be
because M2 limited the growth of degraders. In the future, it will be
necessary to study the changes in functional microbial community in different
mowing treatments using high throughput sequencing. Different from M1/2, M1
significantly increased N-alkyl/methoxyl C, which was recalcitrant C, and
it was relatively enriched in topsoil when O-alkyl or di-O-alkyl C was
prone to oxidation. In terms of the stability of SOM, M1 was the optimized
mowing management practice.
Relationship among net N mineralization, microbes and chemical
compositions of SOM
In natural grasslands, SOM mainly comes from plant litter, roots and soil
microbial cellular residues, ca. 2–15 % of which was constituted by the
N-containing compounds, such as amino acids, amino sugar, pyrimidines,
purines or porphyrin (Mathers et al., 2007). Therefore, close correlations
among the four C functional groups, net N mineralization and MBC were also
observed, and MBC was significantly related to both C functional groups and
net N mineralization (Tables 8 and 9), which suggested that microorganisms
were the driving force of soil C and N turnover in the semiarid natural
grassland. Li et al. (2017) reported that fungi might have played a more
important role than bacteria in the N mineralization in the semiarid Inner Mongolia
grassland, as fungi could better bear the drought and poor available nutrient
conditions (Andresen et al., 2014; Mariotte et al., 2015). We also found
that the correlation between C functional groups and MBC was consistent with
that between C functional groups and net N mineralization. Stevenson et
al. (2016) concluded that soils relatively rich in N should also be
relatively rich in alkyl C, and the chemical composition of SOM significantly
influenced soil N mineralization.
Similar to forest soil, recalcitrant C (alkyl C and aromatic C) also
accounted for a large proportion of the SOC in the grassland soil. In our
study, alkyl and aromatic C accounted for 40.9–47.1 % of all functional
C. It was reported that fungi played the key role in the decomposition of
soil organic N in the forest ecosystem (Boeni et al., 2014; Li et al., 2013),
which indicated that fungi might also be critically important for the
degradation of organic N in the grassland ecosystem. Li et al. (2017)
reported that mowing once a year increased fungal abundance and diversity,
while a higher mowing frequency decreased them. The increased fungal
communities were characterized by the function of mineralizing SOM and
activating nutrients. In the semiarid grassland, the
contents of soil rapidly available N and available P were very low, and
mycorrhizal fungi were richer in M1. Northup et al. (1998) found a mechanism
by which plant productivity could be sustained through mycorrhizal fungi by
investigating plant–soil–microbe interaction. M1/2 and M1 improved herbage
productivity and thus increased net SOC content mainly by increasing
recalcitrant C, and further increased microbial community diversity and
dominant microbial community abundance. In turn, the increased microbial
community enhanced the labile SOM degradation and the humification of SOM to
make the chemical composition of SOM more stable, and this agreed with the
studies conducted by Baumann et al. (2013) and Zhang et al. (2015). Thus, the
relationship between chemical composition of SOM, SOM mineralization and the
microbial community would give us a better understanding of the stability of
soil C and N pools.