Effects of crabs on greenhouse gas emissions, soil nutrients, and stoichiometry in a subtropical estuarine wetland

Crabs may elicit effects on wetland carbon (C), nitrogen (N), and phosphorus (P) concentrations and associated ecological stoichiometry. In this study, we assessed effects of crabs on carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) emissions; soil C, N, and P concentrations; and stoichiometry in upper and mid-tidal flats of an estuarine wetland in China. The results showed that averaged CO2, CH4, and N2O fluxes were greater in the upper and mid-tidal flats in the presence of crabs, being 46.4, 66.7, and 69.7% and 53.6, 143, and 73.1% greater than control, respectively. Mixed model analyses showed overall positive relationships between wetland soil CO2 CH4 and N2O emissions (F = 4.65, P = 0.033; F = 42.42, P = 0.042 and F = 10.2, P = 0.0018, respectively) in the presence of crabs, taking into account season, flooding intensity, and plot effects. This may be related to the direct effects of respiration and the indirect effects of feeding, excretion, and disturbance of soil on microorganisms and/or plant roots. There were no effects of crabs on total C or N concentrations, whereas decreased soil total P concentrations, especially in the upper-tidal flats (P = 0.04). Crab presence was positively associated with soil C/P and N/P ratios (P < 0.0001 and P < 0.0001, respectively), taking into account season, flooding intensity, and plot effects. In the upper and mid-tidal flats, soil CO2 emissions were negatively correlated with total soil C; CH4 emissions were positively correlated with ratios of C/N and C/P; and N2O emissions were positively correlated with N content. In general, global warming potential (GWP) of the upper-tidal flats in the presence of crabs increased by 138% compared with the absence of crabs, and GWP of the mid-tidal flats in the presence of crabs increased by 99.3% compared with the absence of crabs. Global warming and associated flooding rise in several coastal wetland areas are favoring benthic fauna number enhancement, and this in turn increases GWP of overall gas emissions further contributing to future warming rise.


Introduction
The global average temperature increased by 0.58°C during the period between 1880 and 2012 (IPCC 2014), and although many factors contributed to this rise in temperature, increasing concentrations of the key greenhouse gases (carbon dioxide: CO 2 ; methane: CH 4 ; and, nitrous oxide: N 2 O) in the atmosphere contributed about 80% to global warming (IPCC 2014). Globally, wetlands account for 5-8% of the land surface area ), yet despite this relatively small area, they represent important sources and sinks of greenhouse gases (Altor and Mitsch 2008;Chen et al. 2010;Bridgham et al. 2013), due to high levels of primary productivity and soil carbon (C) and nutrient storage (Mander and Shirmohammadi 2008). Wetlands are increasingly important in the retention and filtration of excess reactive niN from surrounding soils and may produce N 2 O by nitrification and denitrification processes (Burgin and Groffman 2012;Cui et al. 2013). Moreover, CH 4 emissions are non-linearly increased with temperature (Peltola et al. 2015) that may elicit a positive feedback on climate change in wetland.
Recent studies of soil C, N, and P have explored the relative balance in concentrations of different elements and turnover processes (Elser et al. 2000) as part of the concept of ecological stoichiometry, which is based on the relationships between ecosystem structure and function, stoichiometry and use of C/N/P in different ecosystem compartments (Elser et al. 2000;Sterner and Elser 2002;Sardans et al. 2012). The relationship between ecosystem energy and multielemental balance described using ratios (Elser et al. 2000) has become a focus on the study of elemental ecological stoichiometry (Elser et al. 2000;Tessier and Raynal 2003;Allen and Gillooly 2009). While spatial variability in ecological stoichiometry and implications for plant growth and ecosystem function have been studied (Han et al. 2005), the effects of benthic organisms on soil nutrient stoichiometry and greenhouse gas emissions are unknown, particularly in estuarine wetlands. The influence of factors at the land-sea interface, periodic tidal inundation, and the effects of benthic fauna on soil and plants is known to affect soil C, N, and P cycling in estuarine wetlands (Liu et al. 2014). Therefore, quantification of nutrient ecological stoichiometry in these ecosystems may increase understanding of drivers and mechanisms of their ecological function and service provision.
As secondary producers in estuarine wetlands, benthic fauna (Stahl et al. 2014) is essential for a healthy ecosystem (Vermeiren and Sheaves 2015), where their bioturbation (Suren et al. 2011), excretion and other physiological activities, and biological irrigation effects impact biogeochemical cycling processes in wetlands (Hu et al. 2016), especially for soil C and N (Liu et al. 2005). Studies have shown that benthic species diversity is greater in the intertidal zone than in the upper mid-tidal zone (Wang et al. 2011;Liu et al. 2015), mainly due to flooding frequency and associated changes in salinity. For example, flooding of the mid-tidal zone brings rapid influxes of nutrients beneficial to benthic organisms and higher levels of salinity that support the growth of algae and other organisms, leading to increased abundance of crabs .
While the effects of benthic fauna on C and N cycling in wetlands have been studied (Mereta et al. 2013), effects on C, N, and P stoichiometry remain unclear. Crabs (Arthropoda: Brachyura) are commonly occurring zoobenthos in wetlands, and they may directly affect soil properties, such as acidity, aeration, and texture, through burrowing activity that subsequently affects C, N, and P cycling (Simioni et al. 2014). For example, CO 2 emissions from burrows in which the crab Upogebia yokoyai was present were greater than in the absence of the crab (Sasaki et al. 2014). CO 2 fluxes at the soil-atmosphere interface were greater in burrows containing crabs (Kristensen et al. 2008), and crab activity in upper tidal flats may promote the production and release of N 2 O as a result of digging behavior, excretion, and sediment resuspension . Other types of benthos have been shown to elicit CH 4 by nutrient-mediated changes to the microbial cycle and coupling between the benthic upper food chain and this can lead to increased CH 4 emissions (Figueiredobarros et al. 2009). Predicted rises in sea levels due to greenhouse gas-meditated global warming are the intensity predicted to lead to increases in the intensity and frequency of flooding that may result in greater crab activity in wetlands, along with associated changes in soil properties and production and emissions of greenhouse gases Wang et al. 2011;Liu et al. 2015). Therefore, it is important to increase the understanding of the role of crabs in the emission of greenhouse gases so as to pormote mitigation. The impacts of crab activities on soil stoichiometry and C, N, and P concentrations and further on greenhouse gas emissions in tidal marshes remain, however, to be studied.
We hypothesized the following: (1) crabs presence increases organic matter and aeration in soil, with associated increases in CO 2 and N 2 O emissions and decreases in CH 4 emissions; (2) crab presence changes physicochemical properties with effects on CO 2 ,CH 4, and N 2 O emissions; and (3) crab presence increases soil C and nutrient concentrations and leads to lower soil C/N and C/P ratios and stimulates emissions of greenhouse gases. To test these hypotheses and to solve these gaps, (1) we determine the effects of the presence of crabs on soil CO 2 , CH 4 , and N 2 O emissions in the upper and middle-tidal flats; (2) we measured the impact of crab presence on soil nutrient levels and their stoichiometry; and (3) we evaluated the relationship between greenhouse gas emissions and soil nutrient concentrations and associated stoichiometry in the presence of crabs.

Study sites
The study sites were located in a tidal Spartina alterniflora wetland in the Minjiang River estuary, China (Fig. 1), where there is a subtropical monsoon climate, with a mean annual temperature of 19.9°C and a mean annual precipitation of 1380 mm. Precipitation is concentrated between March and September, with bimodal peaks in June in the rainy season and August in the typhoon period (Liu et al. 2006). The dominant plant species in the study area included Kandelia candel, Casuarina equisetifolia, Phragmites australis, Cyperus malaccensis, Scirpus triqueter, and Spartina alterniflora.
Three replicates of 1 × 1 m sampling plots were established in four experimental treatments that comprised combinations of presence (P) and absence of crabs (D, control) at upper (G) and mid-tidal (Z, control) flats (GP, GD, ZP, and ZD). Midtidal flat is the high-flooding habitat, which is flooded by intermediate tides ca. 240 d y −1 and is submerged beneath 10-120 cm of water for 0.5-4 h during each tidal inundation (average of 540 h y −1 of inundation). Upper-tidal flat is the intermediate-flooding habitat nearer to coastal line, which is flooded by intermediate tides ca. 220 d y −1 and is submerged beneath 10-100 cm of water for 0.5-3 h during each tidal inundation (average of 385 h y −1 of inundation). All measurements have been performed ± 1 h from low tide. The presence of crabs at the sampling plots was confirmed by the presence of flat holes and chimney burrows (> 4 cm and our in situ investigation showing that these holes were crab holes): in the upper tidal flats, crab burrows were larger in diameter but fewer in number than in the mid-tidal flats (Table S1). To avoid the effects of isolated crabs, control plots (D) were covered with rust-proof barbed wire, polypropylene random (PPR) tubes, and 60-mesh nylon mesh that was installed 70 cm above ground level and 30 cm below ground level.

Greenhouse gas sampling
Greenhouse gas emissions from the plots were collected using static chambers that consisted of an upper chamber and lower base made of opaque PVC (polyvinyl chloride), with diameter and heights of 0.2 × 0.4 m, and 0.2 × 0.1 m, respectively. There was a extraction gas and temperature measurement hole in the top of the upper chamber (Fig. S1). The base of the static chamber used to collect gas samples in the crab-free plots, whereas there was a ≥ 4 cm-hole in the center of the base of the static chamber used in the crab plots to facilitate normal crab activity.
We collected gas samples at low tide in 2015 on January 10 (winter), April 18 (spring), June 27 (summer), and September 13 (autumn) at 09:00, 12:00, and 15:00 h. At each time step, three samples were collected consecutively for 15 min; 40 ml of gas were sampled from the upper chamber and injected into a 50 mL aluminum foil gas sample bag. The temperature in the chamber was recorded during gas collection by the thermometer.
We used gas chromatography to determine CO 2 and CH 4 (Shimadzu GC-2010, Kyoto, Japan) and N 2 O (Shimadzu GC-2014, Kyoto, Japan) concentrations in the gas samples using a Porapak Q stainless steel column (2 m in length, 4 mm OD, 80/100 mesh). A methane conversion furnace, flame ionization detector (FID), and electron capture detector (ECD) were used for the determination of the CO 2 , CH 4 , and N 2 O concentrations, respectively. The operating temperatures of the column, injector, and detector for the determination of CO 2 , CH 4 , and N 2 O concentrations were adjusted to 45, 100, and 280°C; to 70, 200, and 200°C; and to 70, 200, and 320°C, respectively. These temperatures were the optimum values for the different parts of the instrument. The gas chromatograph was calibrated before and after each set of measurements using 503, 1030, and 2980 μL of CO 2 L −1 in He; 1.01, 7.99, and 50.5 μL of CH 4 L −1 in He; and 0.2, 0.6, and 1.0 μL of N 2 O Fig. 1 Location of the study area and sampling sites in southeastern China L −1 in He (CRM/RM information center of China, Beijing) as primary standards (Wang et al. 2015a, b).

CO 2 CH 4 and N 2 O fluxes and the global warming potential (GWP) estimation
The CO 2 CH 4 and N 2 O fluxes were estimated by the following equation: where F is the CO 2 , CH 4 ,or N 2 O flux (mg/μg CO 2 /CH 4 /N 2 O m −2 h −1 ); M is the molecular weight of the gas (44, 16 and 44 g mol −1 for CO 2 , CH 4 and N 2 O, respectively); V is the molar volume of gas in a standard state (22.4 mol −1 ); dc/dt is the variation ratio of CO 2 , CH 4 , and N 2 O concentrations (μmol mol h −1 ); H is the height of the chamber above the water surface (m); and T is the air temperature inside the chamber (°C).
To estimate GWP, CO 2 is typically taken as the reference gas, and a change in the emission of CH 4 or N 2 O is converted into "CO 2 -equivalents" (Hou et al. 2012): where f (kg/hm 2 ) represents the amount of different greenhouse gas emissions during the sampling period. t represents the time of the period. According to Ahmad et al. (2009) the equation used to calculate GWP is: The GWP for CH 4 is 34 (based on a 100-year time horizon and a GWP for CO 2 of 1), and the GWP for N 2 O is 298 (Myhre et al. 2013).

Soil sampling
More than 95% of large benthic fauna is distributed in the upper 10 cm of soil. The diameter of a crab hole is greater at the soil surface and extend to around 10 cm. Therefore, we collected five soil samples from the upper 10 cm of soil of each plot; these samples were mixed to reduce heterogeneity and interference from crab holes , placed in closure pockets, and transported to the laboratory. Impurities, such as stones, wood, and plastic, were removed from the soil, and after air-drying, the samples were passed through a 2-mm sieve prior to measurement of physicochemical soil properties (total C, N, and P concentrations, dissolved organic C, soil exchangeable NH 4 + -N, NO 3 − -N concentrations, soil bulk density, water content, pH, salinity, and temperature). Half of the samples were subsequently passed through a 0.149mm sieve prior to measurement of soil nutrient.
Total C and N (TC and TN, respectively) concentrations were measured using a Vario MAX CN Elemental Analyzer (Elementar Scientific Instruments, Hanau, Germany), and total P (TP) concentration was determined using Mo-Sb colorimetry (Lu 1999;Ruban et al. 1999). Soil DOC was determined by extracting the soils with deionized water (1:5 ratio) and measuring the total C concentration using a TOC-V CPH total C analyzer (Shimadzu Scientific Instruments, Kyoto, Japan). Soil exchangeable NH 4 + -N, NO 3 − -N concentrations were measured using a continuous flow injection analyzer (Skalar Analytical SAN ++ Instruments, Breda, Netherlands).
Bulk density was measured from three 5 × 3-cm cores, while soil water content and soil pH were determined using the drying method and a portable IQ 150 pH meter in a 1:5 ratio of soil/water (IQ Scientific Instruments, Carlsbad, USA), respectively. Salinity and soil temperature in the field were measured using a 2265FS salinity/temperature meter (Spectrum Technologies Inc., Paxinos, USA).

Statistical analyses
We used general mixed models (GLM) in the "nlme" (Pinheiro et al. 2016) R package with the "lme" function, with crab presence, tidal flat (upper or mid to represent flooding intensity), and season as independent fixed categorical variables, plot as independent random factor and CO 2 , CH 4 , and N 2 O soil gas emissions and soil physicochemical properties as continuous dependent variables. When a variable did not follow normal distribution, it was log transformed to reach normal distribution before statistical analyses. We chose the best model for each dependent variable based on the Akaike information criterion, and we used the MuMIn package (Barton 2012) in R to estimate variance explained by the mixed models. Tukey's post hoc tests in the "multcomp" (Hothorn et al. 2013) R package, with the "glht" function, were used to test for differences between the eight combinations of crab presence × season.
Associations between environmental factors and CO 2 , CH 4 , and N 2 O emissions and between C, N, and P concentrations were tested using Pearson correlation analysis in SPSS 20.0 (StatSoft, Inc. Tulsa, USA).
Effects of crab presence and flooding intensity on soil greenhouse gas emissions and soil physicochemical properties were tested using general discriminant analysis (GDA) in Statistica 8.0 (StatSoft, Inc. Tulsa, USA), with a supervised statistical algorithm that derived an optimal separation between groups, which was established a priori by maximizing between-group variance while minimizing within-group variance, to control for effects of season (Raamsdonk et al. 2001).
We used major axis (MA) and standardized major axis (SMA) (SMATR package; http://www.bio.mq.edu.au/ ecology/SMATR) regression to compare the slopes of the regressions of the relationships among the C, N, and P with and without crab.

Results
General effects of crab presence on greenhouse gas emissions CO 2 emissions were globally higher with than without crabs (P < 0.05, Table S2). Average CO 2 emissions from soils in the presence and absence of crabs in the upper tidal flats were 125 ± 14.8 and 67.1 ± 7.67 mg m −2 ·h −1 , respectively (CVs: 23.7 and 22.9%, respectively), and in the mid-tidal flat, average emissions were 128 ± 62.4 and 58.8 ± 20.2 mg m −2 ·h −1 , respectively (CVs: 97.2 and 68.7%, respectively) (Fig. 2). There was no effect of crabs on CO 2 emissions, regardless of degree of tidal inundation (P > 0.05). CO 2 emissions were higher in winter and summer than in spring and autumn (Table S2, Fig.4).
Cumulative CH 4 emissions were significantly higher with than without crabs (Table S2). Mean CH 4 emissions in the upper tidal flats with than without crabs were 0.35 ± 0.22 and 0.10 ± 0.05 mg m −2 ·h −1 , respectively (CVs: 128 and 98%, respectively), and in the mid-tidal flats, mean CH 4 emissions were 0.29 ± 0.26 and − 0.11 ± 0.18 mg m −2 ·h −1 , respectively (CVs: 180 and 320%, respectively) (Fig. 3). There was no effect of crabs on soil CH 4 emissions, regardless of degree of tidal inundation. N 2 O emissions were globally higher with than without crabs (P < 0.05, Table S2). N 2 O emissions in the upper tidal flats were greater in the presence of crabs, especially during the autumn (Fig. 4), and overall, average N 2 O emissions in the tidal flats were significantly greater in the presence of crabs (Table S2, P < 0.05). N 2 O emissions were higher in autumn than in the other three seasons and were higher under high tidal intensity (Table S2, Fig.4).
Under the action of crabs, the CO 2 and N 2 O global warming potential (GWP) of the upper tidal and mid-tidal flats were significantly higher than those without crabs (P < 0.05). GWP of the upper tidal flats in the presence of crabs increased by 138% compared with the absence of crabs, and GWP of the mid-tidal flats in the presence of crabs increased by 99.3% compared with the absence of crabs (Tables S2 and S3).

Crab presence and soil properties
In general, crab presence was associated with greater levels of soil water content, pH. Maximum and minimum soil temperatures were recorded in June and January, respectively; maximum and minimum soil salinities were recorded in January and September, respectively; and maximum and minimum soil pH values were recorded in September and June, respectively. Soil bulk density was greater in the upper tidal flat than Winter Spring Autumn Summer Fig. 2 Effect of presence of crabs on CO 2 emission fluxes from soils of a Spartina alterniflora wetland. Different uppercase letters indicate within season treatment differences and different lowercase letters indicate between season treatment differences (P < 0.05) middle tidal flat. There was no effect of crab presence on seasonal variations in soil temperature, salinity, pH, water content, or salinity, but soil water content was greater with than without crabs (Fig. S2).
There was an overall effect of crab presence on soil TP, but not on soil TC or TN through the year (Table S2, S4). Concentrations of soil TC and TN were greater in the upper tidal flats in the winter and spring than in the upper tidal flats (P < 0.05). Crab presence was associated with greater levels of soil C/P and N/P ratios (Table S2).
Season had a significant influence on the total soil nutrients and their stoichiometric ratios. TN varied greatly for different tidal flats and was related to the frequency of soil flooding (P < 0.05). TC and C/N were significantly influenced by different seasons, different tidal flats, and the presence or absence of crabs. The change of C/P was significant under the combined effect of seasons and crabs and TP in different seasons (P < 0.001). The ratio of soil C/N was stable among the seasons in the upper tidal flats in the presence of crabs, while that of N/P varied; ratios of soil C/N and N/P in the upper tidal flat in the presence of crabs were stable (Fig. 5). Nutrient ratios were greater with than without crabs, and the ratio of soil C/N ratio varied with season (P < 0.01, Table S2).
The y1 and y2 functions represent the linear fitting relationship of total nutrients with and without crab activity, respectively. The R 2 value in the y1 function was significantly higher than the y2 function, indicating that the crab activity helps to improve the correlation between the total soil nutrients (Fig. 6). We observed that none of the slopes of the regression lines of the TN-TC, TP-TC, and TP-TN differed significantly between with and without crabs (P > 0.05, SMA test of common slopes). Concentrations of soil TC, TN, and TP were related (Fig. 6, P < 0.01), but the relationship between TC and TP concentrations was weaker than between TC and TN concentrations. Overall, differences in soil concentrations of TC and TN were more distinct than differences in TP during the year.

Relationship between greenhouse gas emissions and soil properties
Overall, CO 2 emissions were positively correlated with soil temperature (R = 0.35, P < 0.05) and negatively associated with soil bulk density (R = −0.36, P < 0.05), CH 4 and N 2 O emissions were positively correlated with soil temperature (R = 0.60, R = 0.57, P < 0.01, respectively), and N 2 O emissions were positively correlated with soil pH (R = 0.44, P < 0.01) and water content (R = 0.39, P < 0.05). However, the relationships changed in function of the presence of crabs depending on the level of tidal intensity. N 2 O emissions in the upper tidal flat in the presence of crabs were positively correlated with pH (R = 0.82, P < 0.01) and water content (R = 0.59, P < 0.05), negatively correlated with salinity (R = −0.59, P < 0.01) ( Table 1) Fig. 3 Effect of presence of crabs on CH 4 emission fluxes from soils of a Spartina alterniflora wetland. Different uppercase letters indicate within season treatment differences and different lowercase letters indicate between season treatment differences (P < 0.05) 0.86, P < 0.01) and water content (R = 0.70, P < 0.05). Fluxes of CO 2 , CH 4 , and N 2 O were positively correlated with soil temperature (R = 0.84, R = 0.89, R = 0.88, P < 0.01, respectively) and in the presence of crabs (P < 0.005), while N 2 O emission flux was negatively correlated with soil bulk density (R = −0.88, P < 0.01). We found that N 2 O fluxes in the upper tidal flats were positively correlated with soil concentration of TN, TP (R = 0.75, R = 0.79, P < 0.01, respectively), and TC (R = 0.63, P < 0.05) in the presence of crabs and with soil concentration of TC, TN, and TP (R = 0.77, R = 0.83, R = 0.78, P < 0.01, respectively) in the absence of crabs (Table 2). CO 2 emissions in the mid-tidal flats were negatively correlated with soil TC (R = −0.58, P < 0.05) in the presence of crabs and with soil TC concentration and ratio of N:P (R = −0.62, R = −0.68, P < 0.05, respectively) in the absence of crabs. CH 4 emissions were positively correlated with C:N (R = 0.71, P < 0.01) and C:P ratios (R = 0.60, P < 0.05), and N 2 O flux was positively correlated with ratio of C:N (R = 0.77, P < 0.01). Overall CO 2 emissions were negatively correlated with soil concentration of TC (R = −0.37, P < 0.05), while CH 4 flux was positively correlated with C/N and C/P ratios (R = 0.36, R = 0.29, P < 0.05, respectively) and N 2 O flux was positively correlated with soil TN concentration (R = 0.35, P < 0.05). The relationships between soil physicochemical factors on soil TC, TN, and TP and soil stoichiometry are provided in supplementary material (Table S4).

Overall effects of crabs and flooding intensity on greenhouse gas emissions and soil physicochemical properties
The GDA showed that N 2 O emissions, soil TC, C/N ratio, water content, bulk density, and temperature clustered in the four combinations of crab presence/absence and flooding intensity (Tables S5, S6). This GDA detected that the presence of crabs in upper tidal had higher overall effects in the studied variables than in the middle tidal intensity (Fig. S4). Plots in the upper-tidal flats were associated with N 2 O emissions, bulk density, and soil C/P and N/P ratios more with than without crab presence. These great differences due to crab presence were instead not observed in the middle tidal intensity plots (Fig. S4). Plots in the middle tidal flats were correlated with higher soil water content and TP more with than without crab presence (Fig. S4).

Discussion
Effects of crab presence on greenhouse gas emissions CO 2 emissions in wetland soils largely derive from respiration of soil animals and plant roots, and the decomposition of organic C by soil microorganisms (Blagodatsky and Smith 2012). This study showed that CO 2 emissions were greater with than without crabs, possibly due to the direct effects of respiration and the indirect effects feeding, excretion, and disturbance of soil on soil microorganisms and/or plant roots. The burrowing activity of crabs may increase the diffusion of gas and increase soil concentration of O 2 , and this increases soil oxidation capacity. Similar increases in organic C mineralization due to high litter decomposition rates in crab burrows observed in other wetlands (Daleo and Iribarne 2009;Weissberger et al. 2009) depended on the stimulation aerobic bacteria respiration (Liang et al. 2015).
Estuarine wetlands are an important source of CH 4 (Du et al. 2016) and emissions of CH 4 mainly depend on the relative balance its production and oxidation processes. In this study, greater CH 4 emissions were correlated with presence of crabs throughout the year. Consumption of plant and animal litter by crabs is converted to detritus and promotes the accumulation of organic C (Moseman-Valtierra et al. 2011), and this stimulates the growth and reproduction of methanogens (Kammann et al. 2009) leading to increased CH 4 production and emission. Burrowing activities of crabs increase the sediment-water-gas contact interface and the diffusion of inorganic N in sediments from the overlying water bodies and accelerate the rate of ammonia formation (Mereta et al. 2013) and its accumulation in soil. Studies have found that an increase in NH 4 + inhibits methanotrophic activity, and this promotes CH 4 emissions (Hu et al. 2015).
Our results showed that N 2 O fluxes were greater in the presence of crabs. Production of N 2 O was the result of a combination of soil nitrification and denitrification (Harley et al. 2015), which are microbial processes, so factors that affect soil microbial activity also impact soil nitrification and denitrification (Angar et al. 2016). Crab burrowing activity results in the heterogeneous creation of anaerobic and aerobic microsites in soils that are used by a diverse range of microbes; soils surrounding burrows were compacted and anaerobic, while burrows facilitate advection and diffusion of O 2 . Microbial activity in these contrasting soil environments generates N 2 O through nitrification and denitrification . Crabs may affect microbial diversity through feeding on bacteria and fungi, mechanical action on organic matter, propagation of microbial propagules, and changes in nutrient availability that leads to increases in soil N 2 O emissions (Cragg and Bardgett 2001). The O 2 exchange interface between soil and water or the atmosphere is increased in crab burrows. This may lead to increased soil nitrification at the burrow wall, while soil behind this wall may remain anaerobic (Kristensen et al. 2008) and support denitrification leading to greater production and release of N 2 O. Studies have shown that crabs in upper tidal flats may elicit environmental effects.
For example, high densities of crabs at small spatial scales accelerate N 2 O release in flooded tidal flats  and crab fragmentation of surface accelerates litter decomposition with the release of N to the soil environment (Chen et al. 2010). In addition, direct input of excreta by crabs increases soil N concentration (Van Nedervelde et al. 2015) that triggers N 2 O emissions due to the increased abundance and activity of nitrifying and denitrifying soil bacteria (Welsh et al. 2015).
Global warming potential (GWP) of the upper tidal flats increased by 138% with than without crabs, and GWP of the mid-tidal flats increased by 99.3% % with than without crabs (Tables S2, S3). This a very noticeable result because highlight the importance that the direct impacts of global warming on ecosystem processes can in turn exert a powerful feed-back effect on future climate change. In this case, a clear positive    Fig. 6 Linear regression of the effect of crabs on relationships between soil C, N, and P. The y1 and y2 functions represent the linear fitting relationship of total nutrients with and without crab activity, respectively feedback has been detected. Rises in sea levels due to global warming are predicted and under this scenario further increase in the intensity and frequency of flooding are expected that may result in greater crab activity in wetlands, along with associated changes in soil properties Wang et al. 2011;Liu et al. 2015). We have demonstrated that higher crab presence favors the production and emissions of greenhouse gases closing a positive feedback effect.
Effects of crab presence on soil C, N, and P concentrations Soil C, N, and P in wetlands may be affected by multiple factors, such as microbial activity, animal disturbance, decomposition of animals and plants, tidal movement, and human activities (Liu et al. 2014). In this study, the disturbance by crabs was correlated with changes in the concentrations of soil C, N, and P that varied with season. The greatest concentrations of soil C, N, and P were recorded in autumn, correlated with plant senescence and higher inputs of litter that results in larger inputs of food to be consumed and decomposed by the crabs. Overall soil TP concentrations were significantly lower in the presence of crabs than in the control. These results were not consistent with the findings of Mortimer et al. (1999) who did not observe changes in soil TP in the presence of crabs. Animal excreta (feces + urine) drive soil N:P ratios (Sitters et al. 2017). Nevertheless, the greater range in soil TC and TN than TP concentration values throughout the year is consistent with lower mobility and transformation rates of soil P than C and N.

Ecological stoichiometry of soil C, N, and P under the presence of crabs
Concentrations of soil C, N, and P vary spatiotemporally with soil factors, such as animal activity, vegetation, climate, and human interference (Cleveland and Liptzin 2007;Liu et al. 2014). We found that C/N, C/P, and N/P ratios were lowest in summer, and these lower ratios were correlated with higher temperatures in summer when microbial activity is also higher, favoring nutrient use and organic mineralization. The greater release of soil C during the summer reduces its soil concentration and thus favors the drop of its ratios with nutrients, indicating that soil C/nutrient ratios are also controlled by soil C retention and input (Xiao et al. 2014). The decrease in the N/P ratio during the summer is likely due to the slower mineralization rates of P than N, while N responded more rapidly to environmental changes than P (Cleveland and Liptzin 2007). These results confirm the observed nutrient limitation that occurs during S. alterniflora growth in the estuarine wetlands of the Minjiang River (Wang et al. 2015c;Mactavish and Cohen 2017) that lead to greater plant N and Puptake, especially during summer.
The ratios of C/N and N/P were greater with than without crabs, and these ratios were greater in the crab-free upper tidal flats than in the crab-free mid-tidal flats. Thus, the presence of TC soil total C, TN soil total N, soil TP total P, C/N soil C/N ratio, C:P soil C/P ratio, N/P soil N/P ratio, GP upper tidal flat in presence of crabs, GD upper tidal flats in absence of crabs, ZP mid-tidal flats in presence of crabs, and ZP mid-tidal flats in absence of crabs. *P < 0.05, **P < 0.01 crabs was correlated with greater N/P ratios and thus specially with the lower TP concentration. Nutrient ratios and C concentration have important implications for C storage. In this study, we found the ecological stoichiometric ratios (C/N, C/P, and N/P) were positively correlated with C stocks, supporting findings of Wang et al. (2016). The N/P ratio in soils of our study without and with crabs (2.03 and 2.14) was lower than the national wetland N/P ratio (13.6) (Zhang et al. 2016); in general, a lower N/P ratio indicates greater primary productivity capacity, but Nlimitation may become more critical for plant growth in this wetland area under increasing temperatures due to climate change (Wang et al. 2015c).
Effect of presence of crabs on relationships between soil greenhouse gas emissions and soil nutrient and ecological stoichiometry The availability of soil organic C is an important factor for microbial-mediated decomposition processes and CO 2 flux (Vieux et al. 2013). Crabs convert litter into debris through feeding and digestion, and this debris is subsequently incorporated into the soil in the form of excrement; this process increases input to the soil of organic substances and abundance and activity of soil microbes that altogether lead to accelerated C-cycling (Žifčáková et al. 2016), accumulation of organic C in the soil, and increased CO 2 production and emissions (Mehnaz et al. 2018). When oxidation of organic matter is incomplete, as expected under flooding conditions, the production of a series of pyrolytic compounds increases, and this increases the resistance to chemical and biological degradation (Shahbaz et al. 2017). However, crab activity increases aeration and destroys large aggregates, allowing greater decomposition of soil organic matter until complete oxidation occurs, resulting in greater CO 2 formation (Lehmann and Kleber 2015;Murphy et al. 2017).
Consistently, in mid-flooding plots without crabs, we found a positive correlation between CH 4 emission flux and ratios of C/N and C/P, indicating that increasing amounts of organic C substrates with lower nutritional quality favored growth and development of soil methanogens and an increase in CH 4 production and emissions in the absence of crabs. This was not observed in the presence of crabs suggesting the role of crabs in increasing soil aeration and destroying large aggregates.

Other drivers of greenhouse gas emissions from wetland soils
In this study, the CO 2 flux of soil was positively correlated with soil temperature, likely as a result of associated increases in soil microbial abundance and activity (Andrews et al. 2000). Increases in soil temperature to 35°C drive increases in the abundance and activity of soil microorganisms and CO 2 emission flux (Pugh et al. 2018). Soils contain autotrophic and heterotrophic respiring microorganisms (Murdiyarso et al. 2010), and temperatures of c. 15-31°C enhance wetland soil microbial activity and accelerate the rate of degradation of soil organic C, leading to greater releases of CO 2 (Yang et al. 2017).
We found that emissions of CH 4 from the soil were positively correlated with soil temperature, supporting a previous study in a wetland (Yang et al. 2017). The associated increase in soil microbial activity with temperature increases soil O 2 consumption, and this stimulates the growth of methanogens (Zheng et al. 2018), thus observing a positive correlation between soil N 2 O flux and soil temperature consistent with Tian et al. (2015).
As soil temperatures rise, soil microbial nitrification and denitrification and N 2 O emission rate similarly increase (Castaldi 2000). The optimum soil temperature range for nitrification is c. 15-35°C, while that for denitrification is c. 5-75°C (Sun et al. 2010), so it is likely that these processes were active in this study, with greater levels of microbial activity and associated N 2 O emissions from soil. The soil water content, which was 44.2-59.1%, was positively correlated with N 2 O flux, probably because nitrification processes are favored in this range of soil water content (Bramley and White 1989;Szukics et al. 2010). These results support a previous study that showed N 2 O production and emissions increased with increasing soil water content (Di et al. 2014). N 2 O flux was positively correlated with soil pH, consistent with what observed in previous reports (Nkongolo et al. 2008;Duan et al. 2018), associated with the microbial growth (Castellano-Hinojosa et al. 2018). Moreover, the wetland anaerobic environment was rich in ammonium substrate, and then N 2 O flux here is mainly produced by nitrification, and its emissions increase with the increase of soil pH.

Conclusion
This study, at the best of our knowledge, is the first that has studied the shift of gas emissions and nutrient stoichiometry due to the presence of crabs in a subtropical estuarine wetland ecosystem. The presence of crabs mainly accelerates the C and N cycling probably due to increasing aeration, allowing greater decomposition of soil organic matter with greater CO2 formation. However, crab presence, especially in up-tidal sites, decreased soil P concentration, thus rising soil C/P and N/P ratio.
Crab activity (feeding and digestion) also favored the incorporation of litter into soil in the form of excrement; this process increased the input of organic substances to soil and the abundance and activity of soil microbes, both leading to accelerated C-cycling and increased CO2 production and emissions. Moreover, crab burrowing activity resulted in the heterogeneous creation of anaerobic and aerobic micro-sites, where soils surrounding burrows were compacted and anaerobic, while burrows facilitate advection and diffusion of O 2 . This soil environment generates the conditions favorable for CH 4 production and for N 2 O emission. Cumulative CH 4 emissions were significantly higher with than without crabs under both tidal intensities, whereas cumulative CO 2 and N 2 O emissions were also higher with than without crabs, but the impact was significantly higher under up-than under middle-tidal intensity.
The rise of sea level due to global warming may increase the intensity and frequency of flooding, thus promoting the activity of crabs in wetlands, along with associated changes in soil properties, thus inducing a positive feedback effect on global warming.