Search
Author
Title
Vol.
Issue
Year
1st Page

Journal of Animal Science - Animal Production

Carbon flux assessment in cow-calf grazing systems1

 

This article in JAS

  1. Vol. 93 No. 8, p. 4189-4199
     
    Received: Feb 20, 2015
    Accepted: June 01, 2015
    Published: July 24, 2015


    2 Corresponding author(s): marilia.chiavegato@usp.br
 View
 Download
 Share

doi:10.2527/jas.2015-9031
  1. M. B. Chiavegato 2*,
  2. J. E. Rowntree and
  3. W. J. Powers
  1. * Department of Animal Science, University of São Paulo, Av. Pádua Dias, 11, Piracicaba, SP 13418-900, Brazil
     Department of Animal Science, Michigan State University, 424 S. Shaw Lane, East Lansing 48824

Abstract

Greenhouse gas (GHG) fluxes and soil organic carbon (SOC) accumulation in grassland ecosystems are intimately linked to grazing management. This study assessed the carbon equivalent flux (Ceqflux) from 1) an irrigated, heavily stocked, low-density grazing system, 2) a nonirrigated, lightly stocked, high-density grazing system, and 3) a grazing-exclusion pasture site on the basis of the GHG emissions from pasture soils and enteric methane emissions from cows grazing different pasture treatments. Soil organic carbon and total soil nitrogen stocks were measured but not included in Ceqflux determination because of study duration and time needed to observe a change in soil composition. Light- and heavy-stocking systems had 36% and 43% greater Ceqflux than nongrazed pasture sites, respectively (P < 0.01). The largest contributor to increased Ceqflux from grazing systems was enteric CH4 emissions, which represented 15% and 32% of the overall emissions for lightly and heavily stocked grazing systems, respectively. Across years, grazing systems also had increased nitrous oxide (N2O; P < 0.01) and CH4 emissions from pasture soils (P < 0.01) compared with nongrazed pasture sites but, overall, minimally contributed to total emissions. Results indicate no clear difference in Ceqflux between the grazing systems studied when SOC change is not incorporated (P = 0.11). A greater stocking rate potentially increased total SOC stock (P = 0.02), the addition of SOC deeper into the soil horizon (P = 0.01), and soil OM content to 30 cm (P < 0.01). The incorporation of long-term annual carbon sequestration into the determination of Ceqflux could change results and possibly differentiate the grazing systems studied.



INTRODUCTION

Greenhouse gas (GHG) fluxes from grassland ecosystems are intimately linked to grazing management. In grasslands, CO2 is exchanged with the soil and vegetation, N2O is emitted by soils, and CH4 is emitted by microbial activities in the digesta and exchanged with the soil. When CO2 exchange with vegetation is included in net GHG exchange calculation, these ecosystems are often observed as GHG sinks (Allard et al., 2007; Soussana et al., 2007). Similarly, the inclusion of soil organic carbon (SOC) accumulation over time in net GHG exchange accounting might result in grasslands with GHG sink potentials (Liebig et al., 2010).

Grassland management choices to reduce GHG budget may involve important trade-offs. Allard et al. (2007) observed that enteric CH4 emissions expressed as CO2 equivalent strongly affected GHG budget in intensively and extensively managed grasslands (average 70% offset of total CO2 sink activity). Conversely, Soussana et al. (2007) observed that the addition of enteric CH4 and N2O emissions from pasture soils resulted in a relatively small offset of total CO2 sink activity (19% average). Grasslands management affects SOC storage by modifying C inputs to the soil, primarily through root turnover and C allocation between roots and shoots (Ogle et al., 2004). Previous studies reported no effect of grazing on SOC stocks (e.g., Milchunas and Lauenroth, 1993), increase (Wienhold et al., 2001), or decrease (Derner et al., 1997). Factors contributing to net GHG exchange include SOC change, soil-atmosphere N2O flux, enteric CH4 emissions, CO2 emissions associated with N fertilizer production and application, and soil-atmosphere CH4 flux (Robertson et al., 2000; Liebig et al., 2010). This study was performed to assess the net GHG exchange, in terms of carbon equivalent flux (Ceqflux), of grazing systems differing in stocking rate and density. We hypothesized that fewer animals per hectare and greater accumulation of SOC due to longer rest periods produce less Ceqflux per hectare.


MATERIALS AND METHODS

Study Site and Pasture Management

The study was based on previously published data (Chiavegato et al., 2015a,b) and additional measures collected at the Michigan State University Lake City AgBioResearch Center (44°18′N latitude, 85°11′W longitude; elevation 377 m, northwestern Michigan). The first system (SysA) was an irrigated low stocking rate, high stocking density system (1 cow/ha stocking rate and 112,000 kg live weight/ha stocking density; rest period from 60 to 90 d), in which cattle were moved 3 times daily. The objective of this system was to monitor the stock density impact on ecological variables (e.g., soil OM [SOM], water infiltration, and nitrogen content). SysA was based on the hypothesis that high-density systems increase the trample:graze ratio, increasing litter layer and more evenly depositing urine and manure across the landscape. Cow-calf pairs grazed each paddocks 2 to 3 times per year, and no irrigation or fertilization was applied to SysA pasture sites.

The second system (SysB) was a nonirrigated high stocking rate, low stocking density system (2.5 cow/ha stocking rate and 32,700 kg LW/ha stocking density, rest period from 18 to 30 d), in which cattle were moved once daily. This system had an alternative management practice more oriented to animal performance than landscape improvement. In SysB the goal was to increase animal performance by more aggressive defoliation and shorter rest periods to encourage greater forage quality than in SysA. This system was irrigated (K-Line Irrigation, St. Joseph, MI) and received urea fertilization on June 3, 2011 (56 kg/ha of actual urea; approximately 30 d before the start of gas sampling; see dates below). In addition to these 2 systems, a 0.6-ha grazing-exclusion pasture site (GRE) was monitored to account for GHG emissions from nongrazed pastures. The site had not been grazed since the late 1990s (pasture sites were mowed as needed). Chiavegato et al. (2015a) describe further the grazing systems present and quantify pasture soil emissions from the systems for 3 grazing seasons. Enteric methane emissions from the pasture systems are reported by Chiavegato et al. (2015b). These published findings were used as the basis for Ceqflux calculations that follow for each of the farm systems. Soil information is presented in Table 1. Homogeneity of soil types across treatments or replicates is a challenge in field studies. The GRE pasture sites had higher sand content (70%) than SysA (55%) and SysB (53%; Table 1). Soil texture has a strong effect on soil pH, water holding capacity, and permeability, which in turn affect soil temperature (Brady, 1990). As an attempt to overcome the differences in soil type, previous field studies have monitored soil characteristics affected by soil texture and associated with GHG emissions. In this study, there was no difference in soil pH among treatments (Table 1). Chiavegato et al. (2015a) monitored soil water content and temperature to correct pasture-derived GHG emissions.


View Full Table | Close Full ViewTable 1.

Summary of soil characteristics in the study area

 
System1 Size fraction, %
Sand Silt Clay Soil type pH
GRE 70.1 16.0 13.8 Sandy loam 5.9
SysA 55.0 23.5 21.5 Sandy clay loam 6.0
SysB 52.8 26.8 20.4 Sandy clay loam 6.2
1GRE: grazing exclusion; SysA: 1 cow/ha stocking rate and 112,000 kg live weight/ha stocking density; SysB: 2.5 cows/ha stocking rate and 32,700 kg live weight/ha stocking density.

Soil Composition Measures

The sampling dates were June 3 and September 15, 2012, and June 30 and September 28, 2013. Soil was sampled at different depths: 0 to 5, 5 to 10, 10 to 20, and 20 to 30 cm. Three paddocks per system were randomly selected and used as replicates (n = 3). In each paddock, 10 soil subsamples were randomly collected at each depth, composited per paddock, and analyzed for SOM and SOC and total soil nitrogen (TSN) contents. Additionally, 2 samples were collected per replicate for soil bulk density (BD) assessment.

Soil Bulk Density Determination.

Bulk density soil samples were collected from each of the 3 replicates approximately 20 d postgrazing in 2012 and 2013. Soil BD samples were collected with a 7.6 cm diam. by 7.5 cm high brass ring, avoiding disturbance of soil structure, and then were dried at 105°C to constant weight. Bulk density was calculated by dividing the dry weight by the soil core volume (Blake and Hartge, 1986). Soil BD in the 0 to 7.5 cm topsoil was used to calculate SOC stock at 0 to 5 and 5 to 10 cm depths. Soil organic carbon stock from 10 to 20 cm was calculated on the basis of BD from 10 to 17.5 cm depth, and SOC stock at 20 to 30 cm was calculated from BD at 20 to 27.5 cm depth.

Soil Organic Matter Determination.

Samples were collected at the following depths for SOM analyses in 2012 and 2013: 0 to 5, 5 to 10, 10 to 20, and 20 to 30 cm. Soil samples were dried at 65°C to a constant weight then sent to the Michigan State University Soil and Plant Nutrient Laboratory for SOM determination. Soil OM was determined by wet digestion and colorimetry (Schulte and Hopkins, 1996).

Soil Organic Carbon and Total Soil Nitrogen Determination.

Soil samples collected in 2012 and 2013 were dried at 65°C to a constant weight and then sent to Michigan State University Great Lakes Bioenergy Research Center Laboratory for analysis of C and N. Soil organic carbon and TSN contents were determined by an Elemental Combustion System (ECS 4010 CHNSO Analyzer, Costech, Valencia, CA). The ECS uses combustion and gas chromatography with a thermal conductivity detector and helium as the carrier gas to determine nitrogen gas and CO2. Testing for the presence of inorganic C in the soils of the study area concluded that inorganic forms of C were not present (data not shown); thus, total C represents SOC. The carbon:nitrogen ratio was calculated for 0 to 30 cm depths. Soil organic carbon and TSN stocks were calculated on the basis of soil layers of fixed depth (Eq. [1]). However, given that we observed high variability in BD between years and among treatments, we corrected SOC and TSN values for a fixed mass of soil, as suggested by Ellert et al. (2002). Equations [2]–[4] use SOC as an example for calculations, but the same equations were used to assess TSN stocks. This approach includes the selection of a reference soil mass (Mref), which is the smallest soil mass at the prescribed depth from all sampling sites. Here Mref determines the soil mass to be subtracted from the deepest core segment (excess mass of soil, Mex).

The SOC stock calculation based on soil layers of fixed volume iswhere SOCFD is soil organic carbon stock to a fixed depth (Mg/ha), Ci is organic carbon concentration at depth i (mg C/g dry soil), BDi is the bulk density of soil at depth i (g/m3), and Li is the length of the depth i (cm).

The determination of soil mass in each depth iswhere Msoil is the mass of soil to a fixed depth (Mg/ha), BDi is bulk density of soil at depth i (g/m3), and Li is the length of the depth i (cm).

The determination of excess soil mass in each depth iswhere Mex is mass of excess soil (Mg/ha), Msoil is the mass of soil to a fixed depth (Mg/ha), and Mref is the lowest soil mass selected from all sampling sites and depths (Mg/ha).

The determination of soil organic carbon stock to a fixed soil mass iswhere SOCFM is the SOC stock for a fixed mass of Mref, Mex is the mass of excess soil (Mg/ha), and Cdl is the organic carbon concentration at the deepest depth (mg C/g dry soil).

Accumulation of SOC and TSN stocks in soil occurs slowly as a result of management. For that reason, grassland C sequestration potential is evident only in long-term monitoring studies (Schipper et al., 2010). The goal of this study was not to understand SOC or TSN stock changes during the studied years (or C sequestration potential of the grazing systems) but to identify any possible short-term differences in SOC and TSN stocks as a result of management given that pasture sites were managed similarly for 10 yr before the study period.

Carbon Equivalent Flux Calculation

Net GHG exchange (NGHGE) of a managed grassland ecosystem was calculated aswhere NEE is the net ecosystem exchange of CO2 that includes emissions from soil and plant respiration, FCH4 is the CH4 flux from soil, and FN2O is N2O flux from the soil (Chapin et al., 2002; Soussana et al., 2007). To obtain the net GHG exchange in terms of Ceqflux, CH4 and N2O emissions were added to CO2 emissions using the global warming potential (GWP) of each of these gases at the 100-yr time horizon (Intergovernmental Panel on Climate Change, 2007; GWPN2O = 298 and GWPCH4 = 25) as follows:where FCO2 is the C equivalent flux of CO2 from the soil, FCH4soil is the C equivalent flux of CH4 from the soil, FCH4cows is the C equivalent flux of enteric CH4 from the cows, and FN2O is the C equivalent flux of N2O from the soil. Greenhouse gas emissions from pasture soils were collected with the static chamber methodology adapted from Livingston and Hutchinson (1995) and de Klein and Harvey (2013). Static chambers were randomly placed on paddocks (n = 3), and gas samples were collected once daily for 14 d postgrazing. A complete description of the static chamber methodology used to measure FCO2, FCH4cows, and FN2O is available in Chiavegato et al. (2015a). Soussana et al. (2007) used the eddy covariance technique, which allowed them to include CO2 lost by plant and animal respiration in the calculation. In this study, we used the static chamber methodology to monitor pasture-derived GHG fluxes (Chiavegato et al., 2015a), and therefore, FCO2 does not include CO2 lost by plant and animal respiration. Enteric CH4 from the calves was not included in the calculation. Because nondigestible C (from 25% to 40% of the intake depending on herbage digestibility) is returned to the pasture mainly as feces (Soussana et al., 2007), no differentiation was made between manure-derived emissions and soil-derived emissions. Soil emissions sampling occurred postgrazing; hence, any emissions from feces or urine decomposition were assumed to be accounted for in the soil term.

Unlike the calculations of Soussana et al. (2007) and Chapin et al. (2002), who accounted for the C lost from plant biomass export, our calculations were limited to the grazing season. No C loss via herbage cutting and removal from the sampled sites was considered in the above equation. Carbon loss from herbage decomposition on top of the soil is assumed to be included in CO2 and CH4 emissions from the soil, SOM, and SOC content. No addition of C into the systems via organic fertilization occurred, and C leaching from pasture soils was considered negligible.

To allow summation of GHG fluxes from soil and cows and determination of Ceqflux, FCH4cows (originally in g CH4·cow−1·d−1) was converted to an area basis (g CH4·ha−1·d−1) using the stocking rates of each system: SysA = 1 cow/ha and SysB = 2.5 cows/ha. Because only grazing season measures were made, Ceqflux is shown as daily average flux rather than an annual flux.

Change in SOC and TSN stocks was not included in the Ceqflux determination because monitoring occurred for only a 2-yr period, which is not considered long enough to detect accurate stock changes (Schuman et al., 2002). However, we consider SOC and TSN total stocks in our discussion of Ceqflux to illustrate the importance of looking at different pools when assessing GHG emissions from grazing systems.

Statistical Analysis

Soil organic carbon and TSN stock data were analyzed as a completely randomized design. Statistical analyses were performed using SAS software (version 9.2; SAS Inst. Inc., Cary, NC). Paddocks were considered experimental units and were treated as the random term, and the compressed term year × period was considered a repeated measure. Means are pooled by year and period. All tests were performed with 95% confidence (α = 0.05).

The Ceqflux data were analyzed as a completely randomized design. Paddocks were considered experimental units and were treated as the random term, and the compressed term year × period was considered a repeated measure. When the main effect of year was significant, differences were discussed separately by year. When the main effects of treatment or period were significant, the interaction treatment × period was evaluated, and preplanned comparisons within treatment and period were performed. All tests were performed with 95% confidence (α = 0.05).

The model was as follows:where μ is the overall mean, ρi is the random effect of the ith paddock (ρi ∼ N(0, σ2ρ)), τl is the fixed effect of the lth treatment, γk is the fixed effect of the kth year, λm is the fixed effect of the mth period, τγlk is the interaction between the lth treatment and the kth year, τλlm is the interaction between the lth treatment and the mth period, and ejlkmi is the residual term (ejlkmi ∼ N(0, Σ)).


RESULTS AND DISCUSSION

Soil Bulk Density

Soil BD values were different from 2012 to 2013 (P < 0.01) but did not change from August 1 (P1) and August 28, 2011 (P2); June 3 (P1) and September 15, 2012 (P2); June 30 (P1) and September 28, 2013 (P2)(P = 0.19). Therefore, means were pooled by period (Table 2). Soil BD increased with soil depth, but no treatment effects were observed (Table 2). Although the greater stocking density of SysA could have increased BD though compaction, litter accumulation may have protected grazed soils from compaction (Sanjari et al., 2008), resulting in no BD differences between grazing systems and GRE. Savadogo et al. (2007) studied the effects of grazing intensity on properties of a silty-clay deep soil. Their soils had greater clay content than the soils of the present study. However, in agreement with our results, Savadogo et al. (2007) did not observe effects of grazing intensity on soil BD. Their stocking rate varied from 0 to 8 animal units/ha, but BD at 10 cm depth remained constant at 1.4 g/cm3. On the other hand, Franzluebbers and Stuedemann (2009) did observe an effect of a sandy-clay loam soil management in BD values. The authors compared tillage treatments applied to grazed or ungrazed pasture areas. The BD values obtained were similar to this study’s values (on average, 1.5 g/cm3); however, they verified a reduction in BD values under tillage.


View Full Table | Close Full ViewTable 2.

Soil bulk density in pasture soils grazed under 2 management strategies and nongrazed

 
Soil bulk density,1 g/cm
Item GRE SysA SysB
2012 grazing season
    Soil depth, cm
        0 to 5 1.27 1.20 1.25
        5 to 10 1.27 1.20 1.25
        10 to 20 1.57 1.25 1.35
        20 to 30 1.43 1.47 1.44
    SEM 0.05
    Source of variation
        Treatment 0.11
        Depth <0.01
        Treatment × depth 0.11
2013 grazing season
    Soil depth, cm
        0 to 5 1.46 1.57 1.39
        5 to 10 1.46 1.57 1.39
        10 to 20 1.65 1.58 1.62
        20 to 30 1.65 1.59 1.57
    SEM 0.04
    Source of variation
       Treatment 0.14
       Depth <0.01
       Treatment × depth 0.36
1GRE: grazing exclusion; SysA: 1 cow/ha stocking rate and 112,000 kg live weight/ha stocking density; SysB: 2.5 cows/ ha stocking rate and 32,700 kg live weight/ha stocking density.

Soil Organic Carbon and Total Nitrogen Stocks and OM Content

Year and period effects on SOC stocks were observed (P < 0.01 and P = 0.05, respectively; Table 3), which are likely associated with spatial and temporal variability. Soil organic carbon stocks display high spatial variability, especially in grasslands (Cannell et al., 1999). Previous research has associated the variability with sampling at different depths (Bird et al., 2002), climate (Conant et al., 2001), texture (mainly clay content; Parton et al., 1987), and the lack of evaluation of C distribution within the grazing system (Schuman et al., 1999). The ability to detect change in SOC stocks depends on the time since the original sampling, the spatial homogeneity of the soil, and the intensity of sampling (Schipper et al., 2010). This study did not use grazing simulation; rather, systems were monitored at farm scale. Therefore, finding the most adequate conditions for sampling all different variables (e.g., GHG emissions from pasture soils, enteric CH4 from cows, SOC, and TSN) was difficult. For different reasons (such as farm management and logistics, paddocks sizes, and number of animals and total land needed to accurately monitor the grazing systems) sampled paddocks across treatments (replicates) were different at each year and period, which did not allow spatial homogeneity between soil samples. However, paddocks were managed similarly from 1990s through 2010 (continuous grazing), soil types were not greatly different (Table 1), and soil BD did not vary between systems (Table 2). Therefore, it is possible that the changes in SOM, SOC, and TSN stocks between treatments could be a result of grazing management, and the relative change between treatments may be considered, especially in the top layers. A longer study period is necessary to confirm that the changes observed in SOM, SOC and TSN stocks were indeed a consequence of grazing management.


View Full Table | Close Full ViewTable 3.

Soil organic carbon and total soil nitrogen stocks in pasture soils grazed under 2 management strategies and nongrazed

 
Stock
C:N
Item SOC,1 Mg/ha TSN,2 Mg/ha
Systems3
    GRE 42.0a 3.44a 21.0
    SysA 47.4a 3.95a 18.7
    SysB 63.0b 4.85b 19.4
SEM 3.8 0.2
Source of variation
Treatment <0.01 <0.01 0.06
a,bWithin columns, means without a common superscript differ(P < 0.05).
1Soil organic carbon.
2Total soil nitrogen.
3GRE: grazing exclusion; SysA: 1 cow/ha stocking rate and 112,000 kg live weight/ha stocking density; SysB: 2.5 cows/ha stocking rate and 32,700 kg live weight/ha stocking density.

Table 3 illustrates SOC stock means pooled by year and period. On average, SOC stock was greater for SysB pasture sites, and the difference between GRE and SysA was not significant (63.0, 42.0, and 47.4 Mg C/ha for SysB, GRE, and SysA, respectively; P < 0.01). Similarly, the SysB pasture sites had greater TSN stocks than GRE and SysA (4.85, 3.44, and 3.95 Mg N/ha, for SysB, GRE, and SysA, respectively; P < 0.01). Nitrogen and C are predominantly covalently bonded in SOM (Schipper and Sparling, 2011), which explains why the pattern of TSN accumulation in pasture sites was similar to SOC accumulation (Table 3).

The SysA pasture sites had longer rest periods (60 to 90 d) than SysB pasture sites (18 to 30 d). Nevertheless, the increased SOC stock of SysB pasture sites, observed in the short term (2 yr), could suggest that grazing management of SysB increased SOC stocks at a faster rate than SysA or GRE (P < 0.01; Table 2). Naeth et al. (1991) suggested that grazing, such as that in SysB, reduces litter mass accumulation because animal traffic enhances physical breakdown and incorporation of litter into the soil. Frequent grazing could have stimulated forage and root development and increased soil water content and microbial development, thereby enhancing the rate of decomposition of litter and transfer of C into deeper layers of the soil (Shariff et al., 1994). Root decay, although not measured in this study, was identified as another reason for increased SOC under rotational grazing systems. Intensive defoliation during a single grazing event results in cessation of plant respiration, leading to death of roots within a few hours after grazing to equalize biomass (Sanjari et al., 2008). Defoliation in SysB was intense and more frequent than in SysA. As a result of rest period length, the forage offered in SysA was mature and in reproductive stage. We reported previously that the quality of forage offered in SysA resulted in selective grazing from cow-calf pairs (Chiavegato et al., 2015b). Consequently, the portion of forage mass not ingested was trampled down, possibly resulting in greater litter accumulation on soil surface. The significantly lower SOC stock in SysA and GRE compared with SysB might be the result of immobilization of C in excessive aboveground plant litter due to longer rest periods (SysA) or nongrazing (GRE).

SysB had greater SOM content to 30 cm than SysA or GRE, and SysA and GRE did not differ from each other (4.07%, 3.33%, and 3.22% for SysB, SysA, and GRE, respectively; P < 0.01). Soil OM decreased throughout the soil profile in all treatments (Fig. 1). It is known that SOC constitutes approximately 60% of SOM (Bardgett et al., 2009). Consequently, the differences in SOM content between treatments were similar to the differences observed for SOC stocks. Our data support earlier findings that change in soil C can extend throughout the soil profile (Schipper et al., 2010, 2007). Soils in SysB contained greater SOC content in the 20- to 30-cm layers compared with those in SysA (P = 0.02) and GRE (P = 0.03; Fig. 2). Although this was a short-term monitoring of C and N stocks, our results could imply that SysB pasture sites accumulated greater C content in deeper layers than SysA and GRE. Further long-term research is needed to confirm these differences. Frequent trampling by the cow-calf pairs in SysB resulted in disruption of surface soil crust and soil aggregates, increasing SOM decomposition and SOC incorporation in deeper depths (Liu et al., 2004; Neff et al., 2005). Conant et al. (2001) and Franzluebbers and Stuedemann (2009) reported that changes in SOC stock were closely related to changes in TSN stock. Since C and N are covalently bounded in SOM, TSN stock was also highly stratified with depth and followed SOC and SOM accumulation (Fig. 3). Potential benefits result from coupling between soil C and N changes. For example, the sequestration or loss of 1 Mg C is associated with approximately 100 kg of N gained or lost (Schipper et al., 2010). There was no treatment effect on C:N ratio (Table 2). The relatively high C:N ratio observed in this study suggests that C and N immobilization is the dominant process over mineralization (Du Preez and Snyman, 1993).

Figure 1.
Figure 1.

Soil organic matter in pasture soils grazed with 2 different grazing management strategies and nongrazed pastures sites.GRE: grazing exclusion; SysA: 1 cow ha-1 stocking rate; SysB: 2.5 cows ha-1 stocking rate.

 
Figure 2.
Figure 2.

Soil carbon stock in pasture soils grazed with 2 different grazing management strategies and nongrazed pastures sites. GRE: grazing exclusion; SysA: 1 cow ha-1 stocking rate; SysB: 2.5 cows ha-1 stocking rate.

 
Figure 3.
Figure 3.

Total soil nitrogen stock along the soil profile in pasture soils grazed with 2 different grazing management strategies and nongrazed pastures sites. GRE: grazing exclusion; SysA: 1 cow ha-1 stocking rate; SysB: 2.5 cows ha-1 stocking rate.

 

Daily GHG exchange and Carbon Equivalent Flux

Means are shown separately by year and period for FCO2, FCH4soil, FN2O, FCH4cows, and Ceqflux because of the significant effect of year (P < 0.01) and period (P < 0.01; Table 4). Means were then pooled by year and period to allow for a general comparison and discussion about differences between Ceqflux of grazing systems and nongrazed pasture sites (Table 5). Pooled means cannot be extrapolated to annual GHG exchange because results presented in this study are limited to specific monitoring periods during 3 grazing seasons.


View Full Table | Close Full ViewTable 4.

Daily greenhouse gas exchange from pasture soils and animal and total C equivalent flux from pasture sites managed under 2 different management strategies and nongrazed pasture sites during 3 grazing seasons

 
Soil emissions,1 kg C·ha−1·d−1
Animal emissions (FCH4cows),2 kg C·ha−1·d−1
Total emissions (Ceqflux),3 kg C·ha−1·d−1
FCO2
FN2O
FCH4soil
Item P14 P24 P14 P24 P14 P24 P14 P24 P14 P24
2011 grazing season
    System5
        GRE
        SysA 10.54 6.07a* 1.16 0.80 −0.18 -0.07 11.35 6.77a*
        SysB 9.74 7.64b* 1.19 1.59 −0.21 0.06* 10.69 9.57b
    SEM 0.41 0.32 0.04 0.64
    Source of variation
        Treatment 0.28 0.07 0.25 0.03
        Period <0.01 0.96 0.02 <0.01
        Treatment × period <0.01 0.08 0.04 <0.01
2012 grazing season
    System
        GRE 8.24 9.13 0.11 0.05 0.01a 0.003 0 0 8.38a 9.18a
        SysA 8.04 8.31 0.44 0.08 0.14b 0.08 3.28 2.26 12.06b 10.75a,b
        SysB 7.11 9.26* 0.31 0.19 0.08a 0.07 4.89 3.43 12.17b 12.73b
    SEM 0.50 0.11 0.04 0.63 0.57
    Source of variation
        Treatment 0.43 0.19 <0.01 0.12 <0.01
        Period 0.15 0.09 0.38 0.03 0.97
        Treatment × period 0.07 0.33 0.51 0.68 0.06
2013 grazing season
    System
        GRE 19.96 8.57a* 0.96a -0.88 0.20 -0.17 20.77a 7.71a*
        SysA 19.72 10.75a,b* 4.75b 0.35* 0.23 0.33b 1.93a 1.61a 26.13a,b 13.40b*
        SysB 21.49 14.97b* 3.23b 0.82 0.26 0.35b 3.26b 6.22b 28.13b 22.49c
    SEM 1.36 0.70 0.18 0.84 1.96
    Source of variation
        Treatment <0.01 <0.01 <0.01 0.02 <0.01
        Period <0.01 <0.01 0.78 0.11 <0.01
        Treatment × period 0.04 0.03 0.02 0.05 <0.01
a–cWithin columns, means without a common superscript differ (P < 0.05).
*Within rows, means without a common symbol differ (P < 0.05).
1FCO2: C equivalent flux of CO2 from the soil. FN2O: C equivalent flux of N2O from the soil. FCH4soil: C equivalent flux of CH4 from the soil.
2FCH4cows: C equivalent flux of enteric CH4 from the cows.
3Ceqflux: net greenhouse gas exchange in terms of C equivalent.
4July 7 to August 3, 2011 (P1); August 13 to 26, 2011 (P2); May 18 to June 10, 2012 (P1); August 21 to September 11, 2012 (P2); May 20 to June 10, 2013 (P1); August 26 to September 17, 2013 (P2).
5GRE: grazing exclusion; SysA: 1 cow/ha stocking rate and 112,000 kg LW/ha stocking density; SysB: 2.5 cows/ha stocking rate and 32,700 kg live weight/ha stocking density.

View Full Table | Close Full ViewTable 5.

Daily greenhouse gas exchange from soil and animal managed under 2 different grazing strategies and nongrazed pasture sites averaged for 3 grazing seasons

 
Item Soil emissions,1 kg C·ha−1·d−1
Animal emissions (FCH4cows),2 kg C·ha−1·d−1 Total emissions (Ceqflux),3 kg C·ha−1·d−1
FCO2 FN2O FCH4soil
System4
    GRE 9.87a 0.25a -0.09a 0 8.88a
    SysA 10.03a 1.56b 0.13b 2.09a 13.96b
    SysB 11.47b 1.17b 0.10b 4.91b 15.34b
SEM 0.66 0.32 0.08 1.09 0.74
Source of variation
    Treatment 0.17 <0.01 <0.01 0.02 <0.01
a,bWithin columns, means without a common superscript differ (P < 0.05).
1FCO2: C equivalent flux of CO2 from the soil. FN2O: C equivalent flux of N2O from the soil. FCH4soil: C equivalent flux of CH4 from the soil.
2FCH4cows: C equivalent flux of enteric CH4 from the cows.
3Ceqflux: net greenhouse gas exchange in terms of C equivalent.
4GRE: grazing exclusion; SysA: 1 cow/ha stocking rate and 112,000 kg LW/ha stocking density; SysB: 2.5 cows/ha stocking rate and 32,700 kg live weight/ha stocking density.

Observed FN2O values ranged from 0.25 to 1.56 kg C·ha−1·d−1 (Table 5), within the range reported by others (Machefert et al., 2002; Freibauer et al., 2004; Soussana et al., 2007).

Regarding FCH4soil, sink activity was observed (FCH4soil range was from −0.09 to 0.13 kg C·ha−1·d−1), although Soussana et al. (2007), when monitoring CH4 fluxes throughout the year, obtained greater emissions (0.2 to 1.3 kg C·ha−1·d−1). The lower sink activity observed was associated with the presence of grazers, suggesting that grazing reduces the on-site sink activity for CH4 (Soussana et al., 2007). In fact, the negative mean of FCH4soil (pooled by year and period) in the present study was observed from GRE pasture sites (Table 5). Deposition of excreta by animals is expected to produce CH4 emissions at a very low level (compared with application of organic fertilizers; Jarvis et al., 2001) but may increase N2O emissions (Smith et al., 2001). In the present study, very low FCH4soil was observed, and when differences between treatments were observed, they were due to FCO2, FN2O, or FCH4cows (Table 5). Generally, the greater stocking rate in SysB increased FCH4cows but did not affect FCH4soil and FN2O.

Grazing Systems vs. Nongrazed Pasture Sites.

Generally, grazing systems had greater Ceqflux than GRE pasture sites, except during P2 of 2012, when the difference between SysA and GRE was not significant (Table 4). The increased Ceqflux from grazing systems was expected because FCH4cows was considered to be zero for GRE.

The initial hypothesis was that Ceqflux would be increased in grazing systems not only because of enteric CH4 but also because of manure decomposition in pasture soils. However, during 2012 the difference between grazing systems and GRE was approximately 3 kg C·ha−1·d−1, which approximates FCH4cows. This suggests that during 2012, grazing did not increase GHG flux from the soil. The Ceqflux pooled by treatment during 2013 (average 19.8 kg C·ha−1·d−1) was greater when compared to 2012 (10.3 kg C·ha−1·d−1) and 2011 (9.6 kg C·ha−1·d−1). Compared with 2013, 2012 was relatively dry, with precipitation concentrated over a few days during the grazing season (Chiavegato et al., 2015a). Low soil moisture content could have decreased GHG flux from the soil in all pasture sites during 2012 compared with 2013 (Van den Pol-Van Dasselaar et al., 1999; Bardgett et al., 2009; Eckard et al., 2010). Note that 2011 does not include FCH4cows measures.

During 2013, Ceqflux of grazing systems was greater than GRE Ceqflux (by approximately 8 kg C·ha−1·d−1 during P1 and 11 kg C·ha−1·d−1 during P2; Table 4). However, the difference between grazed and ungrazed systems exceeded the enteric CH4 emissions observed during 2013 (on average, 3.3 kg C·ha−1·d−1). Generally, during 2013 GRE pasture soils had decreased FCO2, FCH4soil, and FN2O compared with grazing systems. Only GRE pasture sites demonstrated N2O and CH4 sink activities during the 2013 grazing season (Table 4). The greater levels of moisture in the soil (compared with 2012) likely increased microbial activity, resulting in increased GHG exchange from pasture soils (Van den Pol-Van Dasselaar et al., 1999; Eckard et al., 2010).

SysA vs. SysB.

During 2011, FCH4cows was not monitored, and Ceqflux represents the addition of FCO2, FCH4soil, and FN2O (Table 4). No treatment differences for FN2O and FCH4soil were observed in 2011. During P2, SysB had greater FCO2 than SysA (7.64 and 6.07 kg C·ha−1·d−1, respectively; P = 0.28), which resulted in greater Ceqflux from SysB pasture sites than SysA during P2. Pooled by treatment, Ceqflux decreased considerably from P1 to P2 (11.2 and 8.2 kg C·ha−1·d−1, for P1 and P2, respectively; P < 0.01). Because there were no consistent differences in FN2O and FCH4soil from P1 to P2, the decrease in Ceqflux is due only to the decrease in FCO2. These results suggest that when FCH4cows is excluded, FCO2 drives Ceqflux in grazed pastures.

During 2012, no system differences were observed for FCO2, FCH4soil, FN2O, or FCH4cows. Consequently, no treatment difference between systems’ Ceqflux was observed (Table 4). Despite the greater stocking rate of SysB (2.5 cows/ha) compared with SysA (1 cow/ha), FCH4cows were not significantly different between grazing systems during P2. We expected greater FCH4cows from SysB because of the greater number of cows per hectare. However, the results suggest that individual SysA cows had relatively high enteric CH4 emissions during 2012 (Table 4).

During 2013, SysB had greater Ceqflux than SysA during P2 (22.49 vs. 13.40 kg C·ha−1·d−1, respectively; P < 0.01). The increased Ceqflux from SysB was a result of greater FCH4cows than SysA during P2 (6.22 vs. 1.61 kg C·ha−1·d−1, respectively; P = 0.02) because SysB did not have increased GHG emissions from soils compared with SysA (Table 4). During P1, again, SysB had greater FCH4cows than SysA (3.26 vs. 1.93 kg C·ha−1·d−1, respectively; P = 0.03). However, Ceqflux was not different between grazing systems (26.13 and 28.13 kg C·ha−1·d−1 for SysA and SysB, respectively; P = 0.13). The decreased FCH4cows in SysA was offset by the numerical increase in FN2O, which increased Ceqflux of SysA. These results suggest that the contribution of enteric CH4 to Ceqflux may not always be the driver of greater GHG emissions. Robertson et al. (2000) showed that half of the total net CO2 equivalent emissions from arable sites was contributed by N2O production. Results indicate that under specific circumstances this concept might apply to grasslands.

Daily Ceqflux Pooled by Year and Period.

Daily means were pooled by year and period to allow an overall view of Ceqflux among systems (Table 5).

Daily Ceqflux from grazing systems was greater than that from nongrazed pasture sites by approximately 5.8 kg C·ha−1·d−1 (P < 0.01). The largest contributor for the greater Ceqflux from grazing systems compared with GRE was FCH4cows. However, pooled across year, grazing systems also had greater FN2O and FCH4soil than GRE. Between grazing systems the difference in Ceqflux was not significant (P = 0.60). The only flux that was different between grazing systems was FCH4cows; SysB had greater FCH4cows than SysA (4.91 vs. 2.09 kg C·ha−1·d−1, respectively; P < 0.01). The increased FCH4cows from SysB was a consequence of greater stocking rate because daily enteric CH4 emissions per cow were not different between systems across years (Chiavegato et al., 2015b). The contribution of FCH4cows in SysB was not large enough to increase Ceqflux.

Conclusion

Grazing systems had greater Ceqflux than nongrazed pasture sites. The largest contributor to increased Ceqflux from grazing systems, when compared with GRE, was enteric CH4 emissions. Nongrazed pasture sites were the only sites with CH4 sink activity. The effect of greater enteric CH4 contribution from SysB due to the greater stocking rate than in SysA was offset by GHG exchange from the soil. In the present study, SysB was potentially increasing SOC stocks at a faster rate than SysA or GRE. Similarly, SOM content on topsoil was greater in SysB compared with SysA and GRE, which suggests faster litter decomposition. Soil organic carbon accumulation on deeper layers (20 to 30 cm) was greater in SysB, which also suggests potential of C sequestration. On the other hand, SysA, with longer rest periods allowing for greater litter accumulation on the topsoil, could result in long-term resilience. For management purposes, grazing should be adaptive, and farm decisions are inherent to grazing management. Both SysA and SysB have opportunities to improve ecosystem services at the farm level, including animal production and food provisioning. Long-term research is needed to confirm SOC stock and SOM decomposition rates of these systems and to allow incorporation of SOC change in Ceqflux determination. The incorporation of C sequestration into the determination of Ceqflux could change results and possibly differentiate the grazing systems studied. Further research on the C sequestration potential of different grazing systems is encouraged.

 

References

Footnotes


Comments
Be the first to comment.



Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.