Emission of C[O.sub.2] from tropical riparian forest soil is controlled by soil temperature, soil water content and depth to water table.
Tropical forests are a major pool of terrestrial carbon, and the cycling of carbon through forest soils is a key part of the global carbon cycle (Cleveland et al. 2011). In tropical forests the riparian zones, where terrestrial and aquatic environments interact, are ecologically important for many inter-related reasons (Gregory et al. 1991). Biogeochemical cycling in riparian areas is important for producing high biodiversity and specific ecosystem functions and, where riparian forest is retained in agriculture-dominated landscapes, it can buffer effects of the surrounding land use on aquatic habitats (Burt and Pinay 2005). Tropical riparian forests are arguably more complex ecosystems than other forests because of the interaction of terrestrial, aquatic and hydrological processes and major seasonal variations in hydrological processes, including discharge, recharge and inundation (Connor et al. 2013). Emission of C[O.sub.2] from soil is a key step in carbon cycling, and research of this flux is important for understanding ecosystem carbon cycling. We are not aware of any field studies that have investigated soil C[O.sub.2] emissions in tropical riparian forests, which deserves attention because riparian areas can account for a large proportion of total land area in the wet tropics, e.g. 21% of the Tully-Murray catchment in Australia (Pert et al. 2010). Because of the unique conditions of these ecosystems, the principal drivers of C[O.sub.2] emissions may differ from those in non-riparian areas.
Soil C[O.sub.2] emissions, whether originating from plant roots or heterotrophs (Andrews et al. 1999; Kuzyakov and Larionova 2005), are influenced by soil conditions. Generally speaking, the main factors affecting emission rates are soil temperature, soil moisture, availability of substrate and soil pH, which are influenced in turn by soil structure, weather and topography (Luo and Zhou 2006). The ability of organisms to respire is impaired if soil is too hot, cold, wet or dry (Bowden et al. 1998; Riese and Butterbach-Bahl 2002). In the humid tropics, temperature and moisture tend to co-vary, which hinders efforts to distinguish the effects of the two factors. Substrate availability is difficult to measure, but increased access to labile carbon sources results in higher C[O.sub.2] emissions (Raich and Potter 1995). Suboptimal soil pH levels limit C[O.sub.2] emissions by preventing biological activity (Luo and Zhou 2006), but it is rare to see pH controlling soil C[O.sub.2] emissions in reasonably healthy ecosystems. Most studies on soil C[O.sub.2] emissions focus on these variables. However, the effects of these or other variables on C[O.sub.2] emissions in tropical riparian forests are unknown, owing to the lack of studies in this part of the landscape.
We hypothesised that C[O.sub.2] emission rates in riparian forest soils are determined by measureable physical and chemical variables, and in this paper we present for the first time measurements of soil respiration and the factors controlling it in a tropical riparian forest, a complex and dynamic system. The work was carried out in order to understand the spatial and seasonal dynamics in soil biological activity at a site where the transfers of nitrogen between the surrounding agricultural landscape and the stream were being studied (Connor et al. 2013). The study combined a field study in which total [C[O.sub.2] emissions were measured over the course of a year, and a laboratory study in which the effects of water content and temperature on hcterotrophic C[O.sub.2] emission from intact soil cores from different depths was studied.
Materials and methods
There are two possible strategies for investigating factors controlling C[O.sub.2] emissions from soils: field studies of naturally occurring variation, or laboratory simulation of conditions. In-situ field measurements of gas emission rates capture realistic conditions and capture both heterotrophic and autotrophic emissions. However, it is difficult to identify and determine the role of each controlling factor. The interactions between factors controlling heterotrophic C[O.sub.2] emissions from soil are best studied in laboratory studies on intact soil cores where controlled conditions allow a clearer assessment of the relationship between emission and other variables (Fang and Moncrieff 2001). Most studies have used one or the other of these techniques, but here we used them in combination, to help distinguish between the various factors controlling soil respiration in the field and, in particular, to isolate the importance of temperature and moisture, the two main determinants of emission rates, which also co-vary in field conditions.
The study area was a riparian zone ~150m by 50m in area within a tropical rainforest along the bank of Behana Creek (145[degrees]50'E, 17[degrees]07'S), in the Wet Tropics Bioregion of Queensland, Australia. The site received 2976 mm of rainfall in 2009, slightly wetter than the average. The area has annual mean maximum and minimum daily temperatures of 29.0[degrees]C and 20.8[degrees]C as recorded at the Cairns Aero recording station (Bureau of Meteorology 2015). Behana Creek is a perennial tributary of the Mulgrave River. In the wet season, October-April, the creek regularly breaches its banks and inundates the riparian zone, especially the low elevation gullies (Connor et al. 2013). The vegetation community of the study area is complex mesophyll vine forest (Tracey 1982). The forest has a closed canopy, with trees up to 30 m tall. There is little ground-layer coverage. The soil has been classified as an Orthic Tenosol (Murtha et al. 1996; Isbell 2002).
Soil C[O.sub.2] emissions were measured over 13 months, between January 2009 and February 2010, at four sites within the study area. The sites, named T1B2, T1B3, T2B3 and T3B3, had elevations of 4.12, 5.40, 5.93 and 3.89 m above mean sea level, respectively (Australian Height Datum) and this difference in elevation creates different depth to water table between sites. At each sampling site, six static chambers were positioned randomly ~1 m apart, giving 24 chambers. Each chamber was a section of PVC pipe, 0.25 m in diameter, 0.1 m long and driven 0.02 m into the soil where vegetation did not occur. The shallow insertion depth minimised damage to existing roots. The chambers were left in-situ (uncapped) for the duration of the study, to minimise disturbance. PVC caps with a rubber seal were fitted to each chamber for sampling gas fluxes using bi-directional syringes and pre-evacuated sample tubes. Capillary tubes were fitted to chambers to prevent pressure change when sampled. Because of the dense forest canopy, there was no vegetation in or around the chambers. The chambers were periodically inundated in the wet season when flooding is common, but they flooded and drained at the same rate as the surrounding soil, due to a 10-mm hole drilled through the chamber at ground level. The holes were sealed with bungs during the emission measurements.
Soil C[O.sub.2] emissions were measured at approximately monthly intervals from 31 January 2009 to 30 December 2009, except for a period during February when sampling was conducted at a high frequency (Fig. 1) during a flood event, and a 3-month period during the late dry season when no sampling was conducted. The change in sampling frequency was designed to help identify seasonal variation by making observations around periods of changing conditions. Sampling took place around midday and it was performed on six chambers simultaneously, so measurement occurred around the same time, in order that diurnal variation would not be significant. Access to chambers was limited by flooding on some days during the wet season. C[O.sub.2] emission was measured by repeatedly sampling the gas in the closed chambers and deriving the rate of emission from the change in [C[O.sub.2] concentration during the first 20 min (Tang et al. 2006). The gas samples were taken by using pre-evacuated vials and then analysed for C[O.sub.2] concentration with a GC-2010 gas chromatograph (Shimadzu, Kyoto) fitted with methaniser and flame ionisation detector, calibrated using certified gas standards. Gas samples were also taken between 20 and 60 min of chamber closure, but the emission rate slowed towards the end of the 60-min period, indicating limiting effect of rising C[O.sub.2] concentration in the chamber then.
Soil temperature and soil water content were measured at each chamber on each sampling day. Soil temperature was measured in the top 0.1m in the soil directly adjacent to the chamber. Volumetric soil water content was measured at 0-0.12 m depth using a HydroSense probe (Campbell Scientific, Logan UT, USA), then converted to water-filled porosity according to bulk density. Soil bulk density was measured by using steel rings 70 mm in diameter and 50 mm in length. Soil cores to 0.1 m depth were taken adjacent to each chamber on each sampling day and the cores from each chamber were combined to form one composite sample for each of the four sites. The soil samples were analysed for total carbon content by combustion, using an elemental analyser (Costech, Milan), and pH (Methods 6B2b and 4A1, respectively, of Rayment and Lyons 2011); and water-extractable carbon (1:5 soil : cold deionised water, shaken for 60 min, centrifuged, filtered at 0.45 [micro]m and analysed for dissolved organic carbon). Depth to water table was measured in piezometers (50-mm PVC pipe, slotted from the bottom to 0.4 m below the ground surface) installed at each sampling site. Daily rainfall was recorded at the study site for the duration of the experiment.
During the wet season in early 2010, concentrations of C[O.sub.2] in soil air, soil water contents and soil temperature were measured at several depths at an intermediate elevation in the study site. A pit was dug and water-content probes and air-sampling ports were inserted at 0.1, 0.5, 1.0 and 2 m depth into the face of the pit's wall. The water-content probes were CS616 time domain reflectomctry probes coupled to a CR1000 data logger (Campbell Scientific). The soil air-sampling ports consisted of air-permeable yet hydrophobic Teflon tube (0.1 m long, 6 mm diameter, protected within a perforated steel tube of 10 mm i.d.) at the depth of interest, attached to stainless steel tubes (3 mm i.d.) that protruded from the soil surface and had a septum at the end. A piezometer (50-mm PVC pipe, slotted for the lower 0.1 m) was also installed at each depth to allow measurement of water table depth and sampling of groundwater. The pit was subsequently back-filled as close as possible to the original bulk density by replacing all of the soil, layer by layer, and left for 2 months before measurements and sampling commenced to allow roots and other organisms to recolonise the disturbed area. Soil air and groundwater samples were taken periodically; results reported here covered a period with no rain (2-9 March 2010) and a period with rain (12-19 March 2010). Soil air samples were analysed on the same equipment as samples from the static chambers.
Laboratory soil incubation
The temperature sensitivity of heterotrophic respiration in soils from the study area was measured by using intact soil cores taken from 0-0.2 and 0.2-0.4m depth at two sites chosen for their difference in elevation, T2B3 (5.9 m elevation) and T3B3 (3.9 m elevation). The cores were sampled with sharpened sections of high-density polyethylene pipe (100 mm diameter) hammered into the soil. Compaction of the soil was minimal. The cores were then excavated from the ground and capped on the underside with steel mesh. At each site, eight soil cores were collected from 0-0.2 m (shallow cores) and 0.2-0.4m (deep cores), leading 32 cores in total. The soil cores were immediately taken to the laboratory, where they were saturated (wetting from the bottom up) and then allowed to drain and dry over 6 months at 24[degrees]C. The temperature treatments were carried out over this drying period.
Soil C[O.sub.2] emission was measured periodically using an LI-8100 portable infrared gas analyser (LI-COR, Lincoln, NE, USA). This experimental design resulted in water content and substrate availability being confounded (both decreased together), but the deeper cores were assumed to have lower substrate availability than the shallower cores, throughout the incubation. Between measurements, the cores were open to the atmosphere at both ends to facilitate drying, but 1 day before measurement of emissions, they were capped at the bottom. At each time of measurement the response of C[O.sub.2] emission to temperature was determined by bringing the cores to four temperatures before measurement (18[degrees]C, 24[degrees]C, 30[degrees]C and 36[degrees]C). The cores were brought to the desired temperature by leaving them in an incubator for 8 h before respiration measurements, starting at the lowest temperature and ending with the highest. The first round of respiration measurements was carried out after the cores had been allowed to drain for 72 h and measurements were then repeated nine times over 6 months. Cores were weighed at each stage of the experiment and were finally oven-dried allowing calculation of soil water content.
For the field experiment, stepwise multiple linear regression analysis was carried out, with C[O.sub.2] emission as the dependent variable and all the other measured variables as the independent variables, using S-PLUS[R] 8.0 for Windows (TIBCO, Palo Alto, CA, USA). In order to fulfil the assumption of normally distributed data, C[O.sub.2] emissions and soil carbon content were logarithm-transformed before analysis. Quadratic terms were included for water-filled porosity and depth to water table because of the curvilinear responses expected. Emission of C[O.sub.2] could be expected to be zero at a water-filled porosity of zero, increase to a maximum at intermediate water-filled porosity, and then decrease to a low value at water-filled porosity = 1. Emission was expected to be low when depth to water table is zero (or negative), to increase rapidly as the water table moves through biologically active surface layers, and then reach an asymptote as it moves through deeper layers with low respiration rates. Linear responses were hypothesised for the other parameters. One-way analysis of variance was carried out to compare C[O.sub.2] emission rates between sites and soil carbon content between sites, to examine whether elevation and depth to water table affected these characteristics.
For the field and laboratory data, the van't Hoff model:
R = [ae.sup.[beta]T] (1)
was used to describe the relationship between respiration rate (R) and temperature (7), where a and [beta] are fitted constants (Lloyd and Taylor 1994). The function was fitted to the data of each group of cores at each time period by iteration using SigmaPlot 10.0 (Systat Software, San Jose, CA, USA). The temperature sensitivity of soil respiration (Q10) of soil [C[O.sub.2] emissions was calculated from the function:
Q10 = [e.sup.10[beta]] (2)
Significance of the regression coefficients, and significance of the difference between coefficients for the two depths, was tested using analysis of covariance (ANCOVA) in SPSS[R] Statistics version 22.
Soil C[O.sub.2] emissions, which ranged from 0.2 to 7.5 [micro]mol [m.sup.-2] [s.sup.-1], tended to be least in the cool dry season and highest in the warm wet season (Fig. 1), correlating significantly with soil temperature and depth to water table (Table 1). Overall, the relationship between emission and temperature was positive (with a Pearson correlation coefficient of 0.51) and the relationship with depth to water table was curvilinear, with emission being greatest at intermediate values of water table depth (Table 1, Fig. 2). The most efficient linear model of emissions as a function of measured parameters, which included temperature (0-0.1 m), depth to water table, soil water-filled porosity (0-0.12 m) and pH (0-0.1 m depth), had [r.sup.2] = 0.355 (Table 1). Relationships between C[O.sub.2] emission and environmental variables were further examined by splitting the dataset in several ways. Emissions were significantly higher (P< 0.001) and more variable in the wet season and lower and less variable in the dry season. Mean ([+ or -]1 standard deviation) dry-season C[O.sub.2] emission was 1.4 [+ or -] 1.0 [micro]mol [m.sup.-2] [s.sup.-1] and wet-season emission was 2.4 [+ or -] 1.4 [micro]mol C[O.sub.2] [m.sup.-2] [s.sup.-1]. During the dry season, measured emission rates were better accounted for by the regression ([r.sup.2] = 0.542; Table 1). During the wet season, less of the variation was explained by the variables ([r.sup.2] = 0.301). Splitting the results by water-filled porosity produced the highest correlations between soil C[O.sub.2] emission and the measured parameters; at water-filled porosity <0.45, 74% of the variation in emission was accounted for by the measured variables, whereas at water-filled porosity >0.45, 25% of the variation in emission was explained (Table 1).
Soil C[O.sub.2] emission rates were highly variable at the temperatures >24[degrees]C. In that temperature range, the lowest rates were associated with shallow water table depth ([less than or equal to] 0.15 m) and the highest rates occurred when the water table depth was >0.15 m (Fig. 2). During the study period, water-filled porosity ranged between 0.12 and 0.87 [m.sup.3] [m.sup.3] soil organic carbon content between 35 and 115 g [kg.sup.-1], pH between 4.8 and 6.0, and water-extractable dissolved organic carbon between 2.7 and 12.5 mg [L.sup.-1]. (Fig. 1). Rainfall, soil temperature and soil water content were all strongly seasonal, with maximum values occurring during the wet season (October-March).
There was no significant relationship between soil [C[O.sub.2] emission and soil total carbon content or pH. Soil total carbon content, which had a mean value of 52 g [kg.sup.-1] and a standard deviation of 8.3 across the study area, did not change over the study period. However, it differed significantly (P < 0.01) between sampling sites. Sites T1B3 and T2B3 (the two highest elevation sites) did not differ significantly in soil carbon content, but all other sites differed significantly from each other.
Soil C[O.sub.2] concentrations, and hence emission rates, were influenced by the interaction between soil water content and water table depth, as shown by monitoring during a period of high soil temperature (mean 27.3[degrees]C, Fig. 3). During a 1-week period of stable water table depth and drying soil (2-9 March 2010), C[O.sub.2] concentrations were highest when soil water content was 0.353 [m.sup.3] [m.sup.-3] (0.1 m) and decreased together with soil water content, which reached 0.292 [m.sup.3] [m.sup.-3] on 9 March 2010. By 12 March 2010, C[O.sub.2] concentrations had risen again due to some moistening of the soil by rainfall (water content 0.330 [m.sup.3] [m.sup.-3]). There was then a large rainfall event (see Fig. 1) and the water table rose to the surface (on 14 March 2010), flushing most of the soil air out of the profile. Subsequently, as the water table fell, the soil dried out (0.363 [m.sup.3] [m.sup.-3] on 17 March 2010 and 0.352 [m.sup.3] [m.sup.-3] on 19 March 2010) and C[O.sub.2] concentrations increased. The profile of C[O.sub.2] concentration indicated that respiration rates were highest at 0.1 m depth.
There were also interactions between soil temperature and water content. On a seasonal time scale, there was a positive relationship between the two parameters (Fig. 1), but additional interactions occurred on shorter time scales. During the wet season, when temperatures were generally high, soil temperatures (at 0.1 m) decreased as the soil became wetter, converging on a temperature of ~26[degrees]C when the water table was at the surface (Fig. 2). This fluctuation of soil temperature in response to water content became weaker with depth, disappearing entirely at 1 m depth, where temperature was a constant 26[degrees]C, even when the water table was deeper (Fig. 2). There were also diurnal fluctuations in soil temperature close to the surface.
All soil cores had low C[O.sub.2] emissions while saturated, with mean values of 1.38[micro]mol [m.sup.-2] [s.sup.-1] for shallow cores and 0.43[micro]mol [m.sup.-2][s.sup.-1] for the deep cores. When soil water content was at field capacity and temperature was 26[degrees]C, the mean emission rates were 6.5 [micro]mol [m.sup.-2] [s.sup.-1] for the shallow cores and 3.3 [micro]mol [m.sup.-2] [s.sup.-1] for the deep cores. At field capacity, water-filled porosity was 0.545 for the shallow cores and 0.488 for the deep cores. At the end of the drying period, mean water-filled porosity was 0.144 for the shallow cores and 0.061 for the deep cores.
Emission of C[O.sub.2] decreased over the course of the incubation, along with soil water content and, presumably, supply of readily available substrate (Fig. 4). Compared with the deep cores, the shallow cores had higher rates of emission and more rapid decreases in emission rates over time. All of the van 't Hoff regressions were significant (P< 0.05) except for the observations of T2B3 while the soil was at field capacity. For both the shallow and deep cores, as soil moisture content decreased from field capacity, the Q10 values (temperature sensitivity) increased (Fig. 5). The regressions of Q10 values against soil water content did not differ significantly (ANCOVA, n=17, P=0.692) between the two depths. Over the entire experiment, a limit in emission rates was not encountered with temperature treatment up to 36[degrees]C, the maximum temperature tested.
In this study, soil C[O.sub.2] emission rates were generally related to soil temperature and water table depth, with variations in the relationship occurring on seasonal and shorter time scales. The relationship with water table depth has not been reported previously for tropical forests on mineral soil. The seasonal patterns in emissions corresponded with previous studies, being highest in the wet season (Tufekcioglu et al. 2001; Kiesc and Butterbach-Bahl 2002; Luo and Zhou 2006) and most spatially variable in the wet season (Kosugi et al. 2007). High wet-season emission rates can be explained by high temperature, availability of water, and supply of substrate due to high root activity and growth rates. All of these characteristics enable greater rates of respiration for both heterotrophs and autotrophs (Nelson et al. 1994; Hanson et al. 2000). However, results of the laboratory incubation suggest that direct effects of temperature and water availability on heterotrophic respiration may be sufficient to explain the seasonal patterns of emission in the field without invoking effects on root growth and respiration. Reponses of root respiration to temperature and water availability might contribute to total C[O.sub.2] emission from the soil in this environment in several different ways. Such responses might correspond with those of heterotrophic respiration, or may be negligible in comparison. Alternatively, root respiration and heterotrophic respiration may respond differently to drivers such as temperature, soil moisture and water table depth, thus contributing to the unknown variation in the models shown in Table I.
Temperature was a primary controller of C[O.sub.2] emission rates in this study, in accord with other studies (Lloyd and Taylor 1994). However, the relationship with temperature was stronger in our study than in a previous one in the region, in non-riparian forest (Kiese and Butterbach-Bahl 2002), possibly because of the larger range of temperatures encountered in our study. The range of C[O.sub.2] emission rates measured was remarkably similar in the two studies, being 0.7-7.4 [micro]mol [m.sup.-2] [s.sup.-1] in the study of Kiese and Butterbach-Bahl (2002). Nevertheless, some low emission rates did occur at otherwise optimal temperatures, which is evidence of limitations by other variables or interactions. Correlation of emissions with temperature was highest when soil was wetter (Table 1), similar to findings in other tropical forests (Sotta et al. 2004). High variability in C[O.sub.2] emissions in wet soil (water-filled porosity >0.45) suggests that respiration 'hotspots' were more common in these conditions; the term is taken from Ohashi et al. (2007) to describe observations where emission rates were double the mean or median value, of which there many in our study. From the incubation experiments, it can be inferred that the optimum temperature for heterotrophic respiration in these soils is >36[degrees]C (the hottest treatment imposed), because emission rates increased up to this temperature. It is worth noting though, that the effects of temperature in this experiment may have been different from those in the field, owing to the relatively rapid and large changes imposed. The increase in temperature sensitivity of respiration (Q10) as the soil cores dried was most likely due to a change in the composition of the microbial community (Andrews et al. 2000) and perhaps the nature of the substrates being decomposed (Gauthier et al. 2010).
The influence of water table depth on C[O.sub.2] emissions has been well studied in peatlands (e.g. Oberbauer et al. 1992; Melling et a!. 2005) and in some temperate forests (Jungkunst et al. 2008; Yu et al. 2008; Huang and Ma 2012). However, we have found no previous studies identifying water table depth as a significant controller of C[O.sub.2] emissions in tropical forest on mineral soil. Low emission rates when water table depth was <0.15 m were presumably due to low respiration (limited by oxygen supply), and low diffusion rates of C[O.sub.2] to the surface. Where and when the water table was below the organic-rich surface soil, those factors were non-limiting. The greater the depth of moist, aerated soil, the greater the respiration by heterotrophs and plant roots, if temperature and other controlling factors are non-limiting (Cui et ai 2005). Our results suggest that the timing and location of emissions is influenced by movement of the groundwater; rising water table levels flush C[O.sub.2] from the soil, a falling water table draws atmospheric air into the soil and leaves the unsaturated soil moist, and lateral movement of the groundwater takes dissolved C[O.sub.2] away from its site of production. Changing soil water content (due to infiltration of rainfall and movement of groundwater) may also have influenced soil C[O.sub.2] concentrations and emissions via its effect on temperature, as found by Sotta et al. (2004). In addition to its effect on C[O.sub.2] emissions, depth to water table could also be expected to influence emissions of C[H.sub.4] and [N.sub.2]O and thus the overall greenhouse gas balance in lowland tropical forests (Yu et al. 2008). The relationship between emissions and the water table in forests on mineral soils is presumably most pronounced where the water table fluctuates frequently through the topsoil. More studies are needed before generalisations can be made, but our results suggest models of soil respiration in lowland tropical forests should take into account depth to water table.
The weak control that soil water content had on emission rates in this study was surprising. Others have reported that emissions increased with water content up to a water-filled porosity of ~0.75 (Kiese and Butterbach-Bahl 2002; Luo and Zhou 2006; Werner et al. 2006). In this study, C[O.sub.2] emissions were only weakly correlated with water-filled porosity (correlation coefficient 0.064), and the relationship was not significant in multiple linear regressions. There are several possible reasons for the discrepancy. Perhaps the effects of other factors (temperature, depth to water table and unmeasured factors) over-rode the effects of soil water content, or perhaps the depth at which soil water content was measured was not the best for determining its effect on C[O.sub.2] emissions. Ohashi et al. (2008) found that C[O.sub.2] emission from a tropical rainforest soil correlated more closely with soil water content at 0.3 and 0.6 m depth than 0.1 m depth. All soil water content measurements for this field study were taken at 0-0.12 m depth.
In contrast to the field observations, C[O.sub.2] emission rates in the laboratory study were closely related to soil water content. Emissions decreased as the soil water content decreased from field capacity, corresponding with the accepted view that field capacity provides optimal conditions for soil respiration (Luo and Zhou 2006). However, field capacity was reached near the start of the experiment, so respiration from decomposition of recently severed roots would also have contributed. The results of the laboratory incubations also help to explain the relationship between emissions and soil temperature and water content in the field. Soil respiration was more sensitive to temperature at low soil water content than high soil water content (Fig. 5). The shallow core soils had more substrate available than the deep core soils and thus higher rates of emission. Despite this difference, the change in the temperature sensitivity (Q10) of C[O.sub.2] emission with water content did not differ significantly between depths. This suggests that, although soil water content and substrate availability were confounded (both decreasing with time of incubation), the change in temperature sensitivity with time was related more closely to soil water content than to substrate availability. Although temperature sensitivity of decomposition of soil organic matter has been linked to recalcitrance of substrates (Davidson and Janssens 2006; Craine et al. 2010), the relationship did not hold at the scale of the soil profile in our study, reflecting the complexity of the relationship (Davidson and Janssens 2006; Kleber and Johnson 2010; Uchida et al. 2010; Hamdi et al. 2013).
Several factors that have been found to influence soil C[O.sub.2] emission in other environments, i.e. soil carbon content, water-extractable carbon content and pH, were weak or insignificant predictors of emissions in this study. This lack of correlation may have been due to the small ranges of those parameters in this study. Other studies have also found soil total carbon content not to be a significant predictor in tropical forest soils (Raich and Potter 1995).
Much of the variance in soil C[O.sub.2] emission in our field study was not accounted for by the measured variables, and a variety of factors may have been responsible. Availability of substrate may have been influenced by quality of organic materials, for example due to recent inputs of easily decomposable material or drying/wetting events, which were not adequately quantified by the measurements of total and dissolved organic carbon. Variability in the availability of nutrients may have been influential, with other studies showing microbial respiration to be influenced by the availability of nitrogen (Riese and Butterbach-Bahl 2002; Cui et al. 2005; Janssens et al. 2010; Fujita etal. 2013) and phosphorus (Cleveland et al. 2011). Root activity may also have been important; the contribution of root respiration to C[O.sub.2] emission is normally large (Tang and Baldocchi 2005) but it was not measured in this study. Therefore, factors affecting root respiration, such as biomass and growth rates and aboveground plant processes, are likely to have been important (Vargas et al. 2010). In addition to such chemical and biological factors, physical factors may have played a role. Rapid fluctuations in water table depth during or just before emission measurements may have caused flow of air into-out of the soil surface, modifying near-surface C[O.sub.2] concentration gradients and emission rates. The rapid fluctuations in water table depth were mostly driven by high rates of recharge and discharge in this hydrologically dynamic environment (Connor et al. 2013), but diurnal fluctuations due to uptake of groundwater by vegetation are also significant in forests with shallow water table (Loheide et al. 2005; Connor et al. 2013). In this environment, it is possible that significant amounts of C[O.sub.2] were removed by dissolution in groundwater and subsequent movement to the stream. Expansion and contraction of surface-soil air due to temperature fluctuations (related to diurnal cycles and soil water content changes) may have had effects on gas movement and concentration gradients similar to the water table fluctuations.
In conclusion, emission of C[O.sub.2] from soil in this seasonally variable, riparian tropical forest was significantly related to soil temperature, soil water content and water table depth. Emissions were highest at moderate soil temperature (24-28[degrees]C), water table depth (0.2-1.5 m) and soil water-filled porosity (0.25-0.79). They were lowest (<0.5 [micro]mol [m.sup.-2] [s.sup.-1]) at low soil temperature (<22[degrees]C) or when the water table was within 0.15m of the surface. Overall, soil temperature and water table depth explained 35.5% of the variation in soil C[O.sub.2] emission. The effects of measured variables were greater in the dry season where temperature was the strongest predictor (explaining most of the variation), corresponding with an increase in laboratory-measured temperature sensitivity of heterotrophic respiration as soil dried. A substantial portion of the variation in soil C[O.sub.2] emissions in the field was not explained by the measured variables, especially during the wet season. Possible factors include availability of nutrients and mass flows of soil gas due to movement of the water table and fluctuations in surface soil temperature. It is clear that modelling of soil respiration in tropical forests with a shallow and fluctuating water table should take into account depth to water table.
This work was funded by the Australian Research Council and the Queensland Department of Environment and Resource Management through grant LP0669439. Bruce and Alli Corcoran kindly allowed us access to their property and Dale Heiner provided much assistance in the field.
Andrews JA, Harrison KG, Matamala R, Schlesinger WH (1999) Separation of root respiration from total soil respiration using carbon-13 labeling during free-air carbon dioxide enrichment (FACE). Soil Science Society of America Journal 63, 1429-1435. doi:10.2136/ sssaj1999.6351429x
Andrews JA, Matamala R, Westover KM, Schlesinger WH (2000) Temperature effects on the diversity of soil heterotrophs and the 5,3C of soil-respired C[O.sub.2]. Soil Biology & Biochemistry 32, 699-706. doi: 10.1016/S0038-0717(99)00206-0
Bowden R, Newkirk K, Rullo G (1998) Carbon dioxide and methane fluxes by a forest soil under laboratory controlled moisture and temperature conditions. Soil Biology & Biochemistry 30, 1591-1597. doi: 10.1016/ S0038-0717(97)00228-9
Bureau of Meteorology (2015) Summary Statistics CAIRNS AERO. Available at: www.bom.gov.au/climate/averages/tables/cw_031011. shtml (accessed February 2016)
Burt TP, Pinay G (2005) Linking hydrology and biogeochemistry in complex landscapes. Progress in Physical Geography 29, 297-316. doi:10.1191/0309133305pp450ra
Cleveland CC, Townsend AR, Taylor P, Alvarez-Clare S, Bustamante MMC, Chuyong G, Dobrowski SZ, Grierson P, Harms KE, Houlton BZ, Marklein A, Parton W, Porder S, Reed SC, Sierra CA, Silver WL, Tanner EVJ, Wieder WR (2011) Relationships among net primary productivity, nutrients and climate in tropical rain forest: a pantropical analysis. Ecology Letters 14, 939-947. doi: 10.1111/j.14610248.2011.01658.x
Connor S, Nelson PN, Armour JD, Henault C (2013) Hydrology of a forested riparian zone in an agricultural landscape of the humid tropics. Agriculture, Ecosystems & Environment 180, 111-122. doi: 10.1016/j.agee.2011.12.006
Craine JM, Fierer N, McLauchlan KK (2010) Widespread coupling between the rate and temperature sensitivity of organic matter decay. Nature Geoscience 3, 854-857. doi:10.1038/ngeo1009
Cui J, Li C, Sun G, Trettin C (2005) Linkage of MIKE SHE to Wetland-DNDC for carbon budgeting and anaerobic biogeochemistry simulation. Biogeochemistry 72, 147-167. doi:10.1007/s10533-004-0367-8
Davidson EA, Janssens 1A (2006) Temperature sensitivity of soil carbon decomposition and feedbacks to climate change. Nature 440, 165-173. doi: 10.1038/nature04514
Fang C, Moncrieff JB (2001) The dependence of soil C[O.sub.2] efflux on temperature. Soil Biology & Biochemistry 33, 155-165. doi: 10.1016/ S0038-0717(00)00125-5
Fujita Y, van Bodegom PM, Venterink HO, Runhaar H, Witte JM (2013) Towards a proper integration of hydrology in predicting soil nitrogen mineralization rates along natural moisture gradients. Soil Biology & Biochemistry 58, 302-312. doi: 10.1016/j.soilbio.2012.12.013
Gauthier A, Amiotte-Suchet P, Nelson PN, Leveque J, Henault C (2010) Dynamics of the water extractable organic carbon pool during mineralisation in soils from coniferous and deciduous forests--an incubation experiment. Plant and Soil 330, 465-479. doi: 10.1007/ s11104-009-0220-x
Gregory SV, Swanson FJ, McKee WA, Cummins KW (1991) An ecosystem perspective of riparian zones. Bioscience 41, 540-551. doi:10.2307/ 1311607
Hamdi S, Moyano F, Sail S, Bemoux M, Chevallier T (2013) Synthesis analysis of the temperature sensitivity of soil respiration from laboratory studies in relation to incubation methods and soil conditions. Soil Biology & Biochemistry 58, 115-126. doi: 10.1016/j.soilbio.2012. 11.012
Hanson PJ, Edwards NT, Garten CT, Andrews JA (2000) Separating root and soil microbial contributions to soil respiration: A review of methods and observations. Biogeochemistry 48, 115-146. doi: 10.1023/A:1006244819642
Huang X, Ma JX (2012) The influence of groundwater on soil respiration rate of Populus euphratica community at lower reaches of Tarim River, Xinjiang, China. Journal of Food Agriculture and Environment 10, 1468-1472.
Isbell R (2002) 'The Australian Soil Classification.' Revised edn (CSIRO Publishing: Melbourne)
Janssens IA, Dieleman W, Luyssaert S, Subke J-A, Reichstein M, Ceulemans R, Ciais P, Dolman AJ, Grace J, Matteucci G, Papale D, Piao SL, Schulze ED, Tang J, Law BE (2010) Reduction of forest soil respiration in response to nitrogen deposition. Nature Geoscience 3, 315-322. doi: 10.1038/ngeo844
Jungkunst HF, Flessa H, Scherber C, Fiedler S (2008) Groundwater level controls C[O.sub.2], [N.sub.2]O and C[H.sub.4] fluxes of three different hydromorphic soil types of a temperate forest ecosystem. Soil Biology <6 Biochemistry 40, 2047-2054. doi: 10.1016/j.soilbio.2008.04.015
Kiese R, Butterbach-Bahl K (2002) [N.sub.2]O and C[O.sub.2] emissions from three different tropical forest sites in the wet tropics of Queensland Australia. Soil Biology & Biochemistry 34, 975-987. doi: 10.1016/S0038-0717(02) 00031-7
Kleber M, Johnson MG (2010) Advances in understanding the molecular structure of soil organic matter: Implications for interactions in the environment. Advances in Agronomy 106, 77 142. doi:10.1016/ S0065-2113(10)06003-7
Kosugi Y, Mitani T, Itoh M, Noguchi S, Tani M, Matsuo N, Takanashi S (2007) Spatial and temporal variation in soil respiration in a Southeast Asian Tropical rainforest. Agricultural and Forest Meteorology 147, 35-47. doi: 10.1016/j.agrformet.2007.06.005
Kuzyakov Y, Larionova AA (2005) Root and rhizomicrobial respiration: A review of approaches to estimate respiration by autotrophic and heterotrophic organisms in soil. Journal of Plant Nutrition and Soil Science 168, 503-520. doi:10.l002/jpln,200421703
Lloyd J, Taylor J (1994) On the temperature dependence of soil respiration. Functional Ecology 8, 315-323. doi: 10.2307/2389824
Loheide SP, Butler JJ, Gorelick SM (2005) Estimation of groundwater consumption by phreatophytes using diurnal water table fluctuations: a saturated-unsaturated flow assessment. Water Resources Research 41, W07030. doi: 10.1029/2005WR003942
Luo Y, Zhou X (2006) 'Soil respiration and the environment.' (Academic Press: Burlington, MA, USA)
Melling L, Hatano R, Goh KJ (2005) Soil C[O.sub.2] flux from three ecosystems in tropical peatland of Sarawak, Malaysia. Tellus 57, 1-11. doi: 10.1111/ j.1600-0889.2005.00129.x
Murtha GG, Cannon MG, Smith CD (1996) Soils of the Babinda-Cairns Area, North Queensland. CSIRO Soils, Divisional Report, No. 123.
Nelson PN, Dictor M-C, Soulas G (1994) Availability of organic carbon in soluble and particle-size fractions from a soil profile. Soil Biology & Biochemistry 26, 1549-1555. doi : 10.1016/003 8-0717(94)90097-3
Oberbauer SF, Gillespie CT, Cheng W, Gebauer R, Sala Serra A, Tenhunen JD (1992) Environmental effects on C[O.sub.2] efflux from riparian tundra in the northern foothills of the Brooks Range, Alaska, U.S.A. Oecologia 92, 568-577. doi: 10.1007/BF00317851
Ohashi M, Kume T, Yamane S, Suzuki M (2007) Hotspots of soil respiration in an Asian tropical rainforest. Geophysical Research Letters 34, L08705. doi:10.1029/2007GL029587
Ohashi M, Kumagai T, Kume T, Gyokusen K, Saitoh T, Suzuki M (2008) Characteristics of soil C[O.sub.2] efflux variability in an aseasonal tropical rainforest in Borneo Island. Biogeochemistry 90, 275-289. doi: 10.1007/ S10533-008-9253-0
Pert PL, Butler JRA, Brodie JE, Bruce C, Honzak M, Kroon FJ, Metcalf D, Mitchell D, Wong G (2010) A catchment based approach to mapping hydrological ecosystem services using riparian habitat: a case study from the Wet Tropics, Australia. Ecological Complexity 7, 378-388. doi: 10.1016/j.ecocom.2010.05.002
Raich J, Potter C (1995) Global patterns of carbon dioxide emissions from soils. Global Biogeochemical Cycles 9, 23-36. doi: 10.1029/ 94GB02723
Rayment GE, Lyons DJ (2011) 'Soil chemical methods Australasia.' (CSIRO Publishing: Melbourne)
Sotta ED, Meir P, Malhi Y, Hodnett M, Grace J (2004) Soil C[O.sub.2] efflux in a tropical forest in the central Amazon. Global Change Biology 10, 601-617. doi: 10.1111/j. 1529-8817.2003.00761.x
Tang J, Baldocchi DD (2005) Spatial-temporal variation in soil respiration in an oak-grass savanna ecosystem in California and its partitioning into autotrophic and heterotrophic components. Biogeochemistry 73, 183-207. doi: 10.1007/s10533-004-5889-6
Tang X, Liu S, Zhou G, Zhang D, Zhou C (2006) Soil-atmospheric exchange of C[O.sub.2], C[H.sub.4], and [N.sub.2]O in three subtropical forest ecosystems in southern China. Global Change Biology 12, 546-560. doi: 10.1111/j. 1365-2486.2006.01109.x
Tracey J (1982) 'Vegetation of the humid tropical region of North Queensland.' (CSIRO Publishing: Melbourne)
Tufekcioglu A, Raich J, Isenhart T, Schultz R (2001) Soil respiration within riparian buffers and adjacent crop fields. Plant and Soil 229, 117-124. doi : 10.1023/A: 1004818422908
Uchida Y, Hunt JE, Barbour MM, Clough TJ, Kelliher FM, Sherlock RR (2010) Soil properties and presence of plants affect the temperature sensitivity of carbon dioxide production by soils. Plant and Soil 337, 375-387. doi: 10.1007/sl 1104-010-0533-9
Vargas R, Baldocchi DD, Allen MF, Bahn M, Black TA, Collins SL, Curiel Yuste J, Hirano T, Jassal RS, Pumpanen J, Tang J (2010) Looking deeper into the soil: biophysical controls and seasonal lags of soil C[O.sub.2] production and efflux. Ecological Applications 20, 1569-1582. doi: 10.1890/09-0693.1
Werner C, Zheng X, Tang J, Xie B, Liu C, Kiese R, Butterbach-Bahl K (2006) [N.sub.2]O C[H.sub.4] and C[O.sub.2] emissions from seasonal tropical rainforests and a rubber plantation in Southwest China. Plant and Soil 289, 335-353. doi: 10.1007/s11104-006-9143-y
Yu K, Faulkner SP, Baldwin MJ (2008) Effect of hydrological conditions on nitrous oxide, methane, and carbon dioxide dynamics in a bottomland hardwood forest and its implication for soil carbon sequestration. Global Change Biology 14, 798-812. doi:10.1111/j.1365-2486.2008.01545.x
I. Goodrich (A), S. Connor (A), M. I. Bird (A), and P. N. Nelson (A,B)
(A) Centre for Tropical Environmental and Sustainability Science, James Cook University, PO Box 6811, Cairns, Qld 4870, Australia.
(B) Corresponding author. Email: firstname.lastname@example.org
Table 1. Multiple regressions of C[O.sub.2] emission (log [micro]mol [m.sup.-2] [s.sup.-1]) against measured variables for the whole dataset (most efficient model) and subsets thereof Cells are blank where the parameter was not a significant contributor to the regression. DWT, Depth to water table; WFP, soil water-filled porosity; TOC, soil total organic carbon content; DOC, soil dissolved (water extractable) organic carbon content Dataset n [r.sup.2] P All data 79 0.355 <0.001 Soil WFP <45% 28 0.744 <0.001 Soil WFP >45% 51 0.254 <0.001 Dry season 24 0.542 <0.001 Wet season 51 0.301 <0.001 Dataset Variables and their coefficients Temperature DWT [DWT.sup.2] ([degrees]C) (m) All data 0.064 0.308 -0.092 Soil WFP <45% 0.087 -0.070 Soil WFP >45% 0.275 0.861 -0.212 Dry season 0.092 0.222 -0.068 Wet season 0.077 0.460 -0.164 Dataset Variables and their coefficients WFP [WFP.sup.2] pH ([m.sup.3] [m.sup.-3]) All data -0.183 -0.107 Soil WFP <45% 4.468 -7.736 Soil WFP >45% Dry season 1.387 1.313 -0.299 Wet season 3.090 2.692 Dataset Variables and their coefficients Intercept TOC DOC (g [kg.sup.-1]) (mg [kg.sup.-1]) All data -0.943 Soil WFP <45% -0.027 -2.142 Soil WFP >45% -5.607 Dry season 1.154 Wet season 0.015 -1.269
|Printer friendly Cite/link Email Feedback|
|Author:||Goodrick, I.; Connor, S.; Bird, M.I.; Nelson, P.N.|
|Date:||May 1, 2016|
|Previous Article:||Effect of rice-husk biochar on selected soil properties in tropical Alfisols.|
|Next Article:||Soil microbial biomass carbon and phosphorus as affected by frequent drying-rewetting.|