Carbon balance of an intensively grazed temperate dairy pasture over four years

Agriculture, Ecosystems and Environment 206 (2015) 10–20

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee

Carbon balance of an intensively grazed temperate dairy pasture over four years S. Rutledge a, * , P.L. Mudge a,b , D.I. Campbell a , S.L. Woodward c, 1, J.P. Goodrich a, 2 , A.M. Wall a , M.U.F. Kirschbaum d, L.A. Schipper a a

School of Science and Environmental Research Institute, University of Waikato, Private Bag 3105, Hamilton, New Zealand Landcare Research Manaaki Whenua Ltd., Private Bag 3127, Hamilton, New Zealand DairyNZ, Private Bag 3221, Hamilton, New Zealand d Landcare Research Manaaki Whenua Ltd., Private Bag 11052, Palmerston North, New Zealand b c

A R T I C L E I N F O

A B S T R A C T

Article history: Received 14 June 2014 Received in revised form 8 March 2015 Accepted 12 March 2015 Available online 22 March 2015

We estimated the net ecosystem carbon (C) balance (NECB) of a temperate pasture in the North Island of New Zealand for four years (2008–2011). The pasture was intensively managed with addition of fertiliser and year-round rotational grazing by dairy cows. Climatic conditions and management practices had a large impact on CO2 exchange, with a severe drought in one year and cultivation in another both causing large short-term (
3 months) net losses of CO2–C (100–200 g C m2). However, CO2 was regained later in both of these years so that on annual timescales, the site was a CO2 sink or CO2 neutral. Management practices such as effluent application and harvesting silage also influenced non-CO2–C fluxes, and had a large impact on annual NECB. Despite these major environmental or management perturbations, both NEP and NECB were relatively constant on annual timescales. It is likely that this apparent resilience of the CO2 and C balance to perturbations was at least partly attributable to the relatively warm temperatures, also in winter, providing good growing conditions year-round (in the absence of major perturbations such as moisture stress). In several instances, the farmer’s decisions aimed at maintaining a constant milk yield between years also appeared to contribute to a relatively stable C balance. Averaged over the full four-year study period, the site was a net sink for both CO2 (NEP = 165  51 g C m2 y1), and total C (NECB = 61  53 g C m2 y1) after non-CO2–C fluxes were accounted for. Annual NEP and NECB values were similar to results collated from other managed temperate grasslands on mineral soils globally, for which average NEP and NECB were 188  44 g C m2 y1 and 44  33 g C m2 y1, respectively. In the global dataset, we noted a general trend for increased C sequestration with increasing NEP, suggesting that it may be possible to meet the dual goal of increased pasture production (thus milk, meat and fiber production) and increasing soil C storage in managed temperate grasslands. Identification of management practices that increase C storage while maintaining or enhancing pasture production requires more standardised reporting between NECB studies, and experiments involving side-by-side comparison of treatment and control plots. ã 2015 Elsevier B.V. All rights reserved.

Keywords: Net ecosystem exchange Carbon dioxide Grassland Pasture Grazed agriculture Eddy covariance Cultivation Carbon balance

1. Introduction Temperate grasslands, which cover 1.25  109 ha globally, are important stores of soil organic carbon (C), containing approximately 12% of the global soil organic carbon (SOC) pool (Watson

* Corresponding author at: Department of Earth and Ocean Sciences, University of Waikato, Private Bag 3105, Hamilton, New Zealand. Tel.: +64 7 838 4055. E-mail address: [email protected] (S. Rutledge). 1 Present address: Seed Force Ltd., P.O. Box 16625, Christchurch 8441, New Zealand. 2 Present address: Global Change Research Group, Department of Biology, San Diego State University, 5500 Campanile Drive, San Diego, CA 92182, USA. http://dx.doi.org/10.1016/j.agee.2015.03.011 0167-8809/ã 2015 Elsevier B.V. All rights reserved.

et al., 2000; Lal, 2004b). Long-term studies in permanent temperate grasslands have found increases, decreases or no change in SOC storage over time (Bellamy et al., 2005; Meersmans et al., 2009; van Wesemael et al., 2010; Schipper et al., 2014), with causes for these different findings often remaining largely unresolved. Recent research has shown that appropriate management of grasslands may aid the sequestration of C in soil organic matter, thereby increasing soil quality, offsetting CO2 emissions to the atmosphere and mitigating climate change (Conant et al., 2001; Lal, 2004b; Soussana et al., 2010). For example, increasing pasture productivity through fertiliser application or irrigation, improved grazing management, introduction of legumes or increasing the duration of grass leys, may increase soil C stocks

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

under grasslands (Conant et al., 2001; Schnabel et al., 2001; Freibauer et al., 2004; Soussana et al., 2004; Smith et al., 2008; Jones, 2010). It is crucial to improve our understanding of the impact of management practices and climatic conditions on soil C dynamics in temperate grasslands, so that the effects of future management and changing climate conditions on these grasslands can be evaluated. Quantifying short-term (<5 years) changes in soil organic matter storage using repeated soil sampling over time is difficult because of the generally large background pool of SOC and high spatial variability (Smith, 2004; Rees et al., 2005). Consequently, changes in SOC storage can generally only be determined if there are at least several years between repeated samplings and/or many replicates. However, short-term (1 year) C balances of ecosystems can be studied using the net ecosystem carbon balance (NECB) method (e.g., Byrne et al., 2007; Smith et al., 2010; Soussana et al., 2010). This technique requires the determination of all inputs and outputs of C to and from an ecosystem to deduce any changes in C stocks in the ecosystem. In natural, unmanaged ecosystems, uptake and release of CO2 (by photosynthesis and respiration) are likely to be the main processes affecting the C balance. In managed agricultural ecosystems, additional non-CO2–C inputs (e.g., feed, applied manure or effluent) and outputs (e.g., harvested biomass or milk) may also be significant contributors to the NECB (Soussana et al., 2010; Zeeman et al., 2010; Mudge et al., 2011). Previous studies of the C balance of managed temperate grasslands have found variable results, and reported sites to be C sinks (e.g., Allard et al., 2007; Byrne et al., 2007), C sources (e.g., Veenendaal et al., 2007; Skinner, 2008), or C-neutral (e.g., Prescher et al., 2010). If there are indeed C gains and losses, it is important to understand what climatic conditions or management practices might be responsible for these trends. To help with answering these questions, we determined the C balance of a grazed pasture in New Zealand over four years. In New Zealand, pastoral agriculture is the dominant land use, with dairy cows grazing about 14% of pastoral land. Dairy farms are typically located on the flattest and most productive land, with the remainder of New Zealand’s pastoral land grazed by sheep, beef cattle, deer and young dairy cattle. Mild climatic conditions allow year-round rotational grazing on dairy farms, without a need for housing cattle over winter, which is a common practice for many Northern Hemisphere pasture systems. The New Zealand dairy industry has undergone substantial intensification in recent decades with increasing stocking rates and use of fertiliser, irrigation and supplemental feed (MacLeod and Moller, 2006; Clark et al., 2007). For flat to rolling land (where most dairy farms are situated), recent research has shown decreases in soil C over the last twenty to thirty years for some soil orders, whereas soil C remained stable for others (Schipper et al., 2014). While long-term changes in soil C stocks have been documented in some pastoral soils, there is still poor understanding of underlying temporal trends, or causes for these changes. This research gap requires more detailed measurement campaigns to identify factors causing changes in soil C, which will aid development of approaches to mitigate C losses, or increase C sequestration. In previous work, we described net ecosystem production (NEP) and net ecosystem carbon balances (NECB) for two years at an intensively managed pasture site on a mineral soil in New Zealand (Mudge et al., 2011). Despite a severe drought in the first year and below-normal temperatures for much of the second year, the site was a net sink for CO2 and total C over the two years. We also measured short-term (
40 day) CO2 fluxes following cultivation at the same site, and found that net C losses ranged from 2 to 6 g C m2 d1 depending on soil moisture content (Rutledge et al., 2014). The current paper extends the research of Mudge et al. (2011) by describing an additional two years of measurements (2008–2011)

11

at the same site. The extended study period allowed us to evaluate the impact of a wider range of typical management practices such grazing by cattle, cutting for silage production and pasture renewal (including cultivation) on NEP and NECB. We compared our findings to a global dataset collated from studies carried out in temperate pasture systems. 2. Methods 2.1. Site description This study took place at Scott Farm, a research dairy farm owned and operated by dairy industry research organisation DairyNZ. The farm is located to the east of Hamilton in the Waikato region on the North Island of New Zealand (374601300 S, 175 220 4100 E, 41 m a.s.l.), and has been used for dairy farming for more than 95 years. Average annual rainfall and air temperature are 1126 mm and 13.8  C, respectively. Soil in the flux footprint of the eddy covariance (EC) system is the Matangi silt loam (Typic Orthic Gley Soil, Hewitt, 1993). This soil has imperfect to poor drainage and a bulk density (0–75 mm) of 780 kg m3. Total porosity of the Ap horizon (0–250 mm) was 0.66 m3 m3, field capacity (10 kPa) 0.54 m3 m3 and permanent wilting point (1500 kPa) 0.25 m m3. Total C and N in the topsoil (0–100 mm) were 7.7% and 0.72%, respectively. 2.2. Farm management Scott Farm is made up of small (0.5 ha) paddocks, which are rotationally grazed year round with an average stocking density of
3 dairy cows/ha. Overall stocking rate and management were similar to commercial dairy farms in the Waikato region, with paddocks generally receiving 150 kg N ha1 per year. In spring, when pasture growth exceeded cow demand, pasture from some of the paddocks was cut for silage. When feed demand exceeded pasture growth (typically autumn, winter, and early spring) supplementary feed was fed to cows in the paddocks. During this study, supplementary feed consisted of the silage cut on the farm in spring, maize grown on the farm over summer and additional palm kernel expeller (PKE) brought in from outside the farm. Management of the pasture and cows at Scott Farm was controlled by DairyNZ researchers and farm staff, and largely outside the control of the C flux measurement team. Over the measurement period, the size of grazing herds ranged from 8 to more than 100 cows per paddock. On average, paddocks in the EC flux footprint (hereafter referred to as ‘EC footprint’) were grazed approximately seven times per year. Because paddocks within the EC footprint were not always grazed at the same time, different paddocks had different pasture covers at any given time. In general, grazing took place with larger herds from mid-2009 onwards. From October 2010 onwards, grazing was more synchronised between paddocks. Before March 2010, plant cover in all paddocks of the EC footprint was perennial ryegrass (Lolium perenne) and white clover (Trifolium repens). In 2010, a change in DairyNZ research trials required renewal of the pasture in some of the paddocks in the EC footprint (Fig. 1). Pasture renewal involved spraying the existing sward with herbicide (22 February), application of effluent, mouldboard ploughing to 200–250 mm depth (22 March), application of lime, power harrowing to prepare the seed bed, and then sowing either a high diversity sward, or a low diversity sward (26 March). See Rutledge et al. (2014) for more details about cultivation and sowing. The high diversity sward comprised of a grass component, either perennial ryegrass, high sugar ryegrass, or tall fescue (Festuca arundinacea), and white clover, plantain (Plantago lanceolata), chicory (Cichorium intybus) and lucerne

12

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

Fig. 1. Map of the part of Scott Farm where measurements of the NECB were made. Percentages give the contribution to the CO2 flux per paddock averaged for the times of ‘good quality CO2 flux data’ during the four year study period. The dashed line contains the area contributing 80% of the CO2 flux. Shaded paddocks underwent pasture renewal (including cultivation) in March 2010.

(Medicago sativa). The low diversity sward consisted of a grass component (either ryegrass, high sugar ryegrass or tall fescue) and white clover. 2.3. NECB measurements The net ecosystem carbon balance (NECB) of the ecosystem was calculated using an equation modified from Eq. (2) in Chapin et al. (2006), and included all the key C inputs and exports. We added several fluxes specific for the managed farm ecosystem, and omitted fluxes assumed small and therefore negligible. It is also important to note that system boundaries for which the NECB was calculated encompassed the soil, pasture and cows in the EC footprint, with the ‘EC footprint’ defined by the ‘source area of good quality CO2 flux data’ (see Fig. 1). Throughout this paper, the term ‘ecosystem’ will be used to refer to system as defined above. The NECB for the ecosystem defined thus was calculated as: NECB = NEP + Ffeed + Feffluent – Fproduct – Fharvest – Fmethane – Fdung – Fresp – Fleach (1) where NEP is net ecosystem production (the difference between uptake of CO2 through photosynthesis (GPP) and release of CO2 through total ecosystem respiration (TER) where positive values of NEP indicate net uptake of CO2 by the ecosystem), Ffeed is the C imported as feed from outside the ecosystem, Feffluent is the C applied to the land as effluent, Fproduct is the C exported as milk or meat, Fharvest is the C permanently exported from the ecosystem in silage, Fmethane is the loss of C to the atmosphere as methane produced from enteric fermentation within cows and from deposited dung, Fdung is the C leaving the ecosystem as dung that is deposited on farm races and in the milking shed, Fresp is the CO2 respired by cows while temporarily outside the EC footprint for milking, Fleach is the loss of C as dissolved organic and inorganic C in groundwater. C loss through erosion was assumed to be negligible, due to the flat topography of the site. In this paper, the terms to the right of NEP in Eq. (1) are collectively referred to as ‘non-CO2–C fluxes’. We assumed that on the annual time scale, the change in biomass in living vegetation and cows was small and that only the change in C storage in soil organic carbon over time (DSOC/Dt) needed to be considered (Ammann et al., 2009), such that NECB  DSOC/Dt. The NECB

measurements expanded on measurements made in 2008 and 2009 described in Mudge et al. (2011). 2.3.1. CO2 flux measurements Net ecosystem production (NEP) was assumed to be equal in size but opposite in sign to net ecosystem exchange of CO2 (so NEP = NEE), which was measured using the eddy covariance technique. Measurements from 2008 and 2009 were described by Mudge et al. (2011), who provide details of the instrumental setup and data analysis. In short, the EC setup consisted of a sonic anemometer (CSAT3, Campbell Scientific Inc., Logan, UT) and an open path infra red gas analyser (IRGA; LI-7500, LI-COR, Lincoln, NE). High frequency (10 Hz) data were sampled and stored by a CR3000 datalogger (Campbell Scientific Inc.). Fluxes were corrected using a Matlab (The Mathworks Inc., Natick, MA, USA) program which performed coordinate rotation (McMillen, 1986), corrected for sonic temperature (Schotanus et al., 1983), applied the frequency response correction (Moore, 1986) and added the density term (WPL term; Webb et al., 1980). Quality control of CO2 fluxes was changed slightly from Mudge et al. (2011) and included removal of fluxes when (i) friction velocity (u*) was below 0.11 ms1 indicating low turbulence conditions; (ii) rain, dew or fog affected the IRGA readings, as indicated by the automated gain control (AGC) signal output by the IRGA deviating from the ‘baseline’; (iii) standard deviation in CO2 density exceeded 15 mg m3 during the day or 30 mg m3 at night; (iv) flux values exceeded a threshold number of standard deviations (night-time = 4, daytime = 3) from the mean flux computed for the appropriate time of day across 20-day moving windows; (v) the IRGA or sonic anemometer reported warnings; (vi) out-of-range values were calculated for the flux (|NEE| > 50 mmol m2 s1); or (vii) no fluxes were calculated due to instrument malfunction or power outage. After filtering, 37.8% of data points remained. The majority of the gaps was caused by unreliable IRGA readings due to rain, fog or dew (64% of gaps) and low turbulence conditions (59% of gaps). Data gaps from different sources overlapped much of the time. Data for 2008 and 2009 had previously been filtered and gapfilled by Mudge et al. (2011), but because the filtering approach was altered in the current analysis, data for 2008 and 2009 were re-analysed using the filtering process outlined above.

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

Gaps in the dataset were filled using the online software described by Reichstein et al. (2005). Gaps in the period when no photosynthesis took place due to the cultivation in 2010 (from spraying of the sward until seedling emergence) were gapfilled in isolation. For each half hour with valid CO2 flux measurements, the contribution of individual paddocks and Scott Farm as a whole to

a)

the measured CO2 flux was estimated using the footprint model by Kormann and Meixner (2001). Measured fluxes could be classified as ‘representative’ of the farm (terminology as used by Göckede et al., 2008), with an average of 84% of the flux originating from within the farm boundary over the whole study period. Land use in the remainder of the ecosystem (beyond the farm boundary) was predominantly pastoral agriculture, but instead of being used for

17-yr normal

Scott Farm

30

-2

-1

K↓ (MJ m d )

40

13

20 10

b)

30-yr normal

Scott Farm

Scott Farm cumulative

1500

200

1000

100

500

0 0.8

0

Cumulative rainfall (mm)

300

c)

0.6

3

-3

VMC (m m )

-1

Rainfall (mm mth )

0

0.4 PWP

Temperature (°C)

0.2 30

air

soil

15 10

-1 -2

30-yr normal air

20

5 15 Carbon flux (g C m d )

d)

25

↓↓

e)

10 5 0 -5 NEP -10 Jan08

Jan09

GPP Jan10 Time

TER Jan11

Jan12

Fig. 2. (a) Shortwave incoming radiation (K#), with daily totals from Scott Farm as gray dots and 15 day running means of daily totals for Scott Farm and 17-year normals from a nearby weather station shown as lines; (b) monthly normal rainfall (1981–2010), monthly rainfall at Scott Farm and cumulative rainfall at Scott Farm (restarted on 1 January each year); (c) Volumetric soil moisture content averaged from measurements at 0.05 and 0.1 m depths with daily averages as gray dots and 15 day running means as a line (PWP refers to permanent wilting point); (d) monthly normal air temperature (1981–2010), and 15 day running mean air and soil temperature measured at 0.05 m at Scott Farm; (e) daily summed net ecosystem productivity (NEP, gray dots), and 15 day running mean NEP, gross primary productivity (GPP) and total ecosystem respiration (TER). The black arrow in panel (e) shows when paddocks in the eddy covariance footprint were sprayed with herbicide to kill existing pasture (22 February). Paddocks were cultivated on 22 March, seed sown on 26 March, and seedlings emerged on 6 April (gray arrow).

14

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

dairy farming, was less intensively grazed by young dairy cattle or horses. 2.3.2. Non-CO2–C fluxes We mainly used the methods of Mudge et al. (2011) to determine non-CO2–C fluxes. However, changes in trial setup and grazing approach in the EC footprint between 2008–2009 and 2010–2011 resulted in changes in availability of pasture and milk production data, requiring us to adapt the methods in 2010 and 2011 for some fluxes. We were also able to improve some of our calculations. Following the method described by Zeeman et al. (2010), we weighed the annual non-CO2–C fluxes from different paddocks in the EC footprint by the annual contribution of that paddock to the CO2 flux as determined by the Kormann and Meixner (2001) footprint model. We also added export of C via transfer of dung by cows to races and the dairy shed. Fluxes from all years (2008–2011) were recalculated using the approach described here, which means that some of the recalculated fluxes for 2008 and 2009 had different values from those reported by Mudge et al. (2011). Detailed descriptions of the approach used to determine the various C imports into and exports out of the ecosystem are given in Supplementary material. 2.3.3. Uncertainty estimates The largest uncertainties in the NECB were generally associated with the CO2 flux measurements. The cumulative effect of random errors on annual sums of NEP was determined following the method described by Dragoni et al. (2007), which considers the contributions from both the measurements of turbulent fluxes (Hollinger and Richardson, 2005) and the gap-filling process. The largest systematic uncertainty was the potential underestimation of night-time respiration under calm conditions, which necessitated the choice of a u* threshold below which half-hourly values were rejected and gap-filled (Wohlfahrt et al., 2008; Yi et al., 2010). The uncertainty associated with the selection of this u* threshold was quantified as half of the range in annual summed NEP calculated using a range of u* thresholds deemed plausible for pastures (u* threshold between 0.07 and 0.15 m s1). Combined uncertainties of annual NEP measurements were 47  64 g C m2 y1 (see Table 2). Following Ammann et al. (2009), the uncertainty range () for the non-CO2–C fluxes was approximated by half the range between the minimum and maximum plausible values. Specific details of the uncertainties associated with each of the non-CO2–C fluxes are outlined in Supplementary materials. For combining errors from various independent sources, Gaussian error propagation rules were followed. When estimating uncertainties in the four-year average NEP and NECB of the current study, uncertainties in all component fluxes were propagated assuming they were uncorrelated between years, except for the systematic uncertainty in NEP associated with the u* threshold selection. When reporting the average annual NEP and NECB from the compiled global dataset, 95% confidence intervals based on the distribution of annual values were used.

Meteorological and soil measurements were averaged over 30minute periods, except for rainfall which was summed over 30 min. 3. Results 3.1. Weather and soil conditions Shortwave incoming radiation (K#) ranged between 1 and 10 MJ m2 d1 in the winter months of June and July to >30 MJ m2 d1 during December and January (Fig. 2a). Annual sums of K# were similar between years (within 2.5%). Although the total annual rainfall varied little between the four years of the study (between 1239 mm and 1483 mm; Table 1), the seasonal rainfall distribution varied considerably (Fig. 2b). Generally in the Waikato region, lowest rainfall occurs in late summer/ early autumn (January–March; Fig. 2b, gray bars). Following the seasonal rainfall distribution, soil moisture for the four year study period was generally lowest in summer and highest in winter and spring (Fig. 2c). The very low rainfall in early 2008 (Table 1; Fig. 2b), combined with high evaporative demand, resulted in the soil moisture content dropping below the permanent wilting point (Fig. 2c), which caused the pasture to senescence. Although some rain fell on three occasions in February and March 2008, the drought did not break until mid-April 2008 after which soil moisture levels increased to approximately 0.50 m3 m3. After the dry conditions in summer 2008, the moisture content stayed above 0.30 m3 m3 for the rest of the study period. The most marked change in soil moisture was experienced in spring 2010. The soil was saturated in September 2010 and as a result of low rainfall in October and November, combined with higher-than-normal shortwave incoming radiation, progressively dried out until mid-December when moisture content values had fallen to around 0.30 m3 m3. Mean annual temperature (MAT) over the study period ranged from 12.8 to 14.3  C (Table 1), with highest mean annual temperatures measured in 2010 and 2011. Comparison of the measured air temperature to the 30-year average for a nearby weather station revealed that January 2008, the summer of 2010– 2011 and May–June 2011 were warmer than normal, whereas the whole of 2009 and August–September 2011 were cooler than normal (Fig. 2d). Soil temperature was similar between the four years, except during the 2007–2008 summer, when the soil was approximately 2.5 degrees warmer than in the other years (presumably due to drought conditions limiting evaporative cooling). Following air temperature, soil temperatures in autumn and the first half of winter in 2009 were markedly lower than during the other three years. 3.2. Seasonal and inter-annual CO2 dynamics Fig. 2e shows gross primary production (GPP), total ecosystem respiration (TER) and net ecosystem production (NEP) over the four-year study period. The different dates of cutting and grazing

2.4. Supporting meteorological and soil measurements Incoming shortwave and quantum radiation (LI200SZ and LI190SZ, LI-COR Inc.) were measured at 3 m, and air temperature and humidity were measured at 2.84 m using a capacitive sensor (HMP45A, Vaisala, Finland). A tipping bucket rain gauge (TB5, Hydrological Services Pty Inc., NSW, Australia) measured rainfall at 0.40 m height. Soil temperature was measured at 0.1 m depth using a thermistor (107, Campbell Scientific Inc.). Water content reflectometers (CS616, Campbell Scientific Inc.) were used to measure volumetric soil moisture at 0.05 m and 0.10 m depth.

Table 1 Main weather characteristics of the four measurement years. Summer rainfall is the sum of rainfall over three months (December of the year before and January and February).

Annual rainfall (mm) Summer rainfall (mm) Mean annual temperature ( C)

2008

2009

2010

2011

1298 128a 13.8

1239 334 12.8

1259 384 14.1

1483 371 14.3

a Because measurements were only started on 17 December 2007, the monthly summed rainfall from a nearby weather station was used for December 2007.

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

15

Table 2 Components (and uncertainties) of the annual CO2 and C budget at Scott Farm 2008–2011 (in g C m2 y1). NECB components for 2008 and 2009 were recalculated from Mudge et al. (2011). The NECB was calculated using Eq (1). See text for calculations and assumptions associated with uncertainty estimates. C flux

2008

2009

2010

2011

Average

GPP TER NEP Ffeed Feffluent Fproduct (milk) Fharvest Fmethane Fdung Fresp Fleach NECB

1931 1741 189.5  54.2 20.2  4.7 0.0 65.9  9.9 12.8  1.8 17.4  3.5 14.6  3.7 9.2  2.9 5.5  5.5 84.3  55.9

2325 2142 183.7  46.6 25.2  5.9 0.0 76.8  15.4 29.6  4.2 20.5  4.1 19.8  5.0 12.4  4.0 5.5  5.5 44.2  50.5

2045 1983 61.0  63.8 34.1  8.0 132.1  54.8 78.2  15.6 0.0 21.3  4.3 20.6  5.2 12.9  4.1 5.5  5.5 88.7  86.4

2481 2254 227.1  52.5 16.2  3.8 0.0 63.9  12.8 107.5  15.2 19.4  3.9 13.7  3.5 8.6  2.8 5.5  5.5 24.7  56.8

2196 2030 165.3  50.5 23.9  2.9 33.0  13.7 71.2  6.8 37.5  4.0 19.7  2.0 17.2  2.2 10.8  1.8 5.5  2.8 60.5  53.2

between years, and varying herd sizes made a direct comparison of CO2 dynamics between years complicated. During 2009 and 2011, the two years with least disturbance, GPP and TER showed a seasonal pattern with highest values in spring and early summer, and lowest values during winter. GPP and TER thereby roughly followed the seasonal variations in their main forcing variables, solar radiation (Fig. 2a) and soil temperature (Fig. 2d), respectively. In 2008, the drought year, both GPP and TER were lowest during the drought, while in 2010 when cultivation took place, GPP reduced to zero during the disturbance, but TER reached a minimum in winter. In general, seasonal changes in TER had a smaller magnitude than seasonal changes in GPP. Fig. 3 shows cumulative NEP for the four years. In 2008, the severe drought at the start of the year caused CO2 losses of 106 g C m2 during the first four months of the year. Once growing conditions improved after the substantial rainfall in mid April 2008 (Fig. 2b), the pasture system was a net sink for CO2 until the end of the year, as shown by the steadily upwards trending cumulative NEP curve in Fig. 3. In summer 2009, dry conditions were less severe than in 2008, but the combination of declining soil moisture, and harvest of pasture (removal of most photosynthetic material) contributed to the site becoming a source of CO2 for seven weeks from February to late March (Mudge et al., 2011). The cooler-than-normal temperatures from May to November 2009 (Fig. 2d) limited net CO2 uptake compared to 2008 (Mudge et al., 2011), as displayed by the only slightly positive slope of the 300 Summer

Autumn

Winter

Spring

250

Cumulative NEP (g C m-2)

200 150

3.3. Full carbon balance

100 50 0 -50

↑

-100 -150 -200

cumulative NEP curve in Fig. 3. The strength of the full-year net CO2 sink was similar in 2008 and 2009, although both GPP and TER were larger in 2009 than in 2008 (Table 2). In March 2010, cultivation of much of the EC footprint (Fig. 1) caused a large net loss of CO2, because photosynthesis was reduced to virtually zero while ecosystem respiration continued (Fig. 2e). Over the 82 days between spraying the old sward with herbicide (22 February) to the first day that the newly establishing sward had again become a net sink for CO2 (15 May), 212 g C m2 was lost. Steady net uptake of CO2 by the establishing sward from mid-May onwards ensured that this C was regained before the end of the calendar year, and on an annual time scale the site was a net sink for CO2 (although uncertainty overlapped with zero: NEP2010 = 61  64 g C m2 y1). From the middle of October 2010 through to the end of 2011, NEP alternated between being positive and negative at regular intervals (Fig. 2e). This switching between source and sink of CO2 was caused by grazing in the EC footprint, which took place with larger herds and was more synchronised between the different paddocks from October 2010 onwards compared to earlier in the study period. For the first half of 2011, GPP exceeded that of previous years and this was matched by higher values for TER during those months (Fig. 2e). Annual GPP, TER and NEP in 2011 were the highest of the four years (Table 2). On an annual time scale, the pasture was a net CO2 sink for three of the four years, with NEP of 190  54, 184  47 and 227  53 g C m2 y1 for 2008, 2009 and 2011, respectively (Fig. 3 and Table 2). In 2010, the site was likely to have been a sink for CO2 as well (NEP of 61  64 g C m2 y1). Annually summed GPP ranged between 1931 and 2481 g C m2 y1, with lowest GPP in 2008, the drought year, and highest in 2011 when the new pastures (including more diverse swards) were fully established (Table 2). A similar pattern was observed in annually summed TER.

↑

2008 2009 2010 2011

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month of year Fig. 3. Cumulative annual net ecosystem production (NEP) from 2008 to 2011. The two arrows relate only to events in 2010, with the black arrow indicating when paddocks in the eddy covariance footprint were sprayed with herbicide to kill existing pasture (22 February), and the gray arrow when seedlings emerged (6 April). Paddocks were cultivated on 22 March and seed sown on 26 March.

Table 2 shows the components of the NECB for all four years. In 2008, 2009 and 2011, NEP was the largest net input of C into the ecosystem, with values ranging from 184  47 to 227  53 to g C m2 y1, which made up between 88 and 93% of total annual C inputs. In 2010, the year the cultivation took place and pasture production stopped during an extended period in autumn, annual NEP was positive (indicating uptake of CO2), but the uncertainty range included zero. Carbon input from effluent in 2010 was 132  55 g C m2 y1, which was 58% of total C inputs, and larger than NEP. Inputs of C through supplemental feeding (Ffeed) ranged from 16 to 34 g C m2 y1 between years, making up 7–15% of total annual C input. Most feed was imported in the cultivation year (2010; 34  8 g C m2 y1) when NEP was lowest of the four years. Ffeed was smallest for 2011, the year with the largest GPP, TER and NEP.

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

3.4. Scott Farm NEP and NECB compared to other managed temperate grasslands To assess how NEP and NECB measured at Scott Farm compared to similar farmed systems globally, we compiled data from NECB studies for managed temperate grasslands reported in the literature. We identified 59 site-years of data, from 21 sites, described in 16 studies (Supplementary Table 1). The majority of sites (18 out of 21) was situated in Europe, with the remaining sites situated in the USA (2) and New Zealand (1). The overview combines intensively and extensively managed systems, with the majority of sites being situated on mineral soils (16 out of 21), and five sites having peaty soils. Where data from a site-year was presented both in the European synthesis study by Soussana et al. (2007) and in a stand-alone paper, data from the latter was used. For this analysis, we assumed that the uncertainty of both NEP and NECB values were 50 g C m2 y1 (cf. Baldocchi, 2003), and therefore site-years with absolute values for NEP or NECB <50 g C m2 y1 were assumed to be sources or sinks not significantly different from zero (indicated by gray areas in Fig. 4). Based on this criterion, farms were net sinks for CO2 in 78% of site-years, but after non-CO2–C fluxes had been accounted for, only 41% of site-years remained as net C sinks, while 27% of site-years were reported as significant sources for C (see also Fig. 5). For the remaining 32% of site-years, the net C balance of farms was not significantly different from zero. On average, annual NECB of sites on mineral soils was 144 g C m2 y1 smaller than annual NEP (Fig. 5 and Supplementary Table 1). 4. Discussion 4.1. Average NEP and NECB at Scott Farm Averaged over the full four year study period, the annual NECB was 61  53 g C m2 y1, which was of similar magnitude to the average NECB of 42  35 g C m2 y1 for the studies on managed temperate grasslands collated in Supplementary Table 1 (excluding Scott Farm and grassland sites over peat). Average NEP of 165  51 g C m2 y1 over the four years was also similar to the average NEP of 190  46 g C m2 y1 for studies presented in Supplementary Table 1 (again excluding Scott Farm and grassland

CO2 source ←

→ CO2 sink

1:1

600 SF (mineral)

400

Mineral

→ C sink

Peat

200

-200 0 0 5

-400 -400

-200

0

200

400

0 41 7 25 10 12

C source ←

0

600

-2 -1

NEP (g C m y )

Fig. 4. Relationship between published values of annual net ecosystem production (NEP) and net ecosystem carbon balance (NECB) for managed temperate grasslands (see Supplementary Table 1 for data). Values for this study (Scott Farm; SF) are black triangles. Assuming confidence intervals of NEP and NECB to be 50 g C m2 y1, values in the gray area are not significantly different from zero. Numbers in the small box in the bottom right-hand corner indicate the percentage of site years of data that occur in each of the corresponding areas of the graph.

sites over peat). Further discussion of annual NEP and NECB from the current study in relation to the literature is presented later. 4.2. Intra and inter-annual variability of NEP and NECB 4.2.1. Drought and cold winter conditions Both climatic conditions and farm management had a large impact on carbon dynamics at Scott Farm. During the first four months of 2008, the severe drought caused large losses of CO2 (
100 g C m2), but warm and wet conditions during winter and spring facilitated good pasture growth and large net CO2 uptake for the rest of the year, so that on an annual basis both NEP and NECB were positive (Mudge et al., 2011). Mudge et al. (2011) also attributed low CO2 uptake during the 2009 winter to the colder than normal conditions which limited pasture growth. Despite annual NEP in 2009 being almost identical to 2008, the 2009 NECB was only about half of that in 2008, possibly due to greater C

0.4 NEP NECB

0.3 Frequency

The C leaving the ecosystem in milk (Fproduct) ranged between 64 and 78 g C m2 y1, and for the first three years of the study period was the largest export of C, making up 47–56% of non-CO2–C exports. In 2011, more than 3.5 times as much silage was exported (Fharvest) than in any of the other years, making it the largest export of C that year (108 g C m2 y1 or 49% of total C exported). In 2010, the cultivation year, no silage was exported from the ecosystem, and less was exported in the drought year, 2008, than in 2009 (Table 2). Exports of C via methane and dung deposition and cow respiration outside the EC footprint were similar between years. This was because these component fluxes mostly depended on the number of cows and lactation period, which were relatively stable from year to year. On an annual basis, the NECB was positive for all four years, indicating that the site was a C sink, although uncertainty ranges included zero in 2009 and 2011. Cumulatively over the four years, total C inputs to ecosystem were almost 890 g C m2, with 74%, 15% and 11% coming from net CO2 fixation (NEP), effluent application and supplemental feed, respectively. Of this C, almost 650 g C m2 was exported again, mostly as milk (44% of total C export) and silage (23% of total C exported). The average NECB for the total study period was 61  53 g C m2 y1 over the four years, indicating that on average the site was a net sink for C.

NECB (g C m-2 y-1)

16

0.2

0.1

0

-500

0

500 -2 -1

Carbon flux (g C m y ) Fig. 5. Frequency distributions of annual net ecosystem production (NEP) and net ecosystem carbon balance (NECB) for temperate grassland sites on mineral soils listed in Supplementary Table 1. NEP and NECB were binned into ranges of 100 g C m2 y1. Positive NEP and NECB indicate a sink for CO2 and C, respectively.

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

17

exports via non-CO2–C pathways in 2009 (e.g., milk and harvested biomass; Table 2).

calculation does highlight how one off management events could have a large impact on annual C balances.

4.2.2. Cultivation The most prominent feature of the annual course of NEP in 2010 was the impact of herbicide application (to kill the existing sward) and cultivation in February–March, which caused the site to become a source of CO2 for 82 days after spraying. There was no obvious increase in total respiration following ploughing, presumably because any increase in soil respiration due to soil disturbance, and the decay of killed plant roots was balanced by reduced autotrophic respiration. The imbalance between TER and GPP, however, increased because GPP dropped to zero while TER remained around 4 g C m2 d1 (Fig. 2e). A number of chamberbased studies have reported short-term tillage-associated peaks in CO2 losses immediately after cultivation (e.g., Reicosky and Lindstrom, 1993; Willems et al., 2011), which have been attributed to degassing of soil pores, and possibly increased microbial respiration as a result of increased aeration and availability of labile C caused by the destruction of aggregates and root death. However, other studies using continuous EC have not observed this ‘flush’ of CO2 emissions following ploughing or tillage (e.g., Baker and Griffis, 2005; Aubinet et al., 2009). Numerous studies have reported increased soil respiration following cultivation of cropped soils (e.g., Baker and Griffis, 2005; Aubinet et al., 2009; Béziat et al., 2009) and frequent cultivation of croplands generally causes a decrease in soil C (Lal, 2004a; Baker and Richie, 2007; Smith et al., 2008). In contrast, the longer-term effects of occasional cultivation on CO2 exchange and soil C dynamics of grasslands have received less attention (Conant et al., 2007; Govaerts et al., 2009). Rutledge et al. (2014) showed that short-term (
40 day) CO2 losses resulting from cultivation events at Scott Farm ranged between approximately 80 and 400 g C m2 depending on the time of year of cultivation and associated soil moisture content. However, the current analysis shows that such temporary source activity (measured here as 212 g C m2 over the
3-month period after cultivation in 2010) did not automatically translate to the site being a source for CO2 on an annual time scale, because conditions for the remainder of the year were conducive to good pasture growth. The pattern of a large CO2 loss early in the year (as a result of cultivation), and accumulation later in the year, is similar to what occurred during 2008 (the year with the drought). This suggests that annual NEP in this pasture system was relatively resilient to major environmental or management perturbations, which was possibly because the relatively warm climate allowed year-round growing conditions (in the absence of major perturbations such as drought conditions). Grasslands in cooler or drier environments are likely to be less resilient to environmental or management perturbations, particularly if the perturbations occur during the growing season. The cultivation event in 2010 also highlighted how different management events were inter-related and the effect that they can have on the NECB. When the site was cultivated, the rate of effluent application was increased and additional feed was imported to help maintain milk production at a similar level. While the disturbance due to cultivation resulted in an initial net loss of CO2 from the system, the large amount of effluent applied contributed to the site being a net C sink in 2010. Had cultivation taken place without effluent application, the site would have likely been a C source in 2010 instead, and the average four-year NECB would have been reduced from 61  53 g C m2 to approximately 27  51 g C m2, a small C sink not distinguishable from zero. We acknowledge that this estimate obtained by simply ignoring effluent application, is probably overly simplistic, because in addition to directly affecting C inputs and TER, effluent application would likely have affected plant growth, and thus NEP as well. However, the

4.2.3. Grazing In managed grasslands, pasture can either be consumed by grazing cows or harvested by cutting. Even though individual grazing events at Scott Farm were sometimes hard to detect because of the mosaic of small paddocks making up the EC footprint, the response to grazing events was still discernible in the NEP time series, especially from October 2010 onwards, when grazing tended to be more synchronised between paddocks and herds were larger (Fig. 2e). Defoliation by rotational grazing or cutting causes a sharp decrease in photosynthetically active plant matter (e.g., Rogiers et al., 2005; Wohlfahrt et al., 2008), the remaining leaves have comparatively low photosynthetic potential (Parsons et al., 2011), and grazing cattle deposit large amounts of fresh dung which is available to be respired, in addition to direct respiration from the cows themselves. These combined effects would have caused the site to be a temporary source for CO2 during and directly after grazing (Fig. 2e). Recurrent periods of sink and source activity are typically observed in grasslands that are frequently cut or grazed (e.g., Jaksic et al., 2006; Wohlfahrt et al., 2008; Hussain et al., 2011). Capturing CO2 losses from grazer respiration with eddy covariance instruments will be highly dependent on wind direction and the location of grazing animals. CO2 losses during cattle grazing tend to occur at rates many times higher than CO2 losses from plant and soil respiration, and therefore if grazing fluxes are missed (or discarded due to filters removing high values), it could lead to bias in CO2 flux estimates (Kirschbaum et al., 2015). The small paddocks (and relatively small herd sizes) at our study site, would have meant that fluxes from cattle respiration would not have dominated overall fluxes as much as they would have if only one large herd had been grazing larger paddocks. This is supported by the generally good agreement between modeled and measured CO2 fluxes at our study site (Kirschbaum et al., 2015). Careful thought needs to be given to the best approach to account for animal respiration under different grazing systems, with coupled EC measurements and modeling being one promising option to explore further (Kirschbaum et al., 2015). 4.2.4. Silage export In 2011, the combination of good growing conditions (high GPP) and DairyNZ experimental requirements, led to more silage being cut in that year than in the other three years (Table 2). This meant that the amount of C exported as silage was substantially higher in 2011 compared to the other three study years. This coincided with 2011 having the largest NEP (227  53 g C m2) but smallest NECB (25  57 g C m2) of the four-year study period (Table 2). Gilmanov et al. (2007) also found the largest NEP in intensively managed hay meadows where grass was predominantly removed by cutting, whereas grazed grasslands were CO2 neutral or a moderate sink. This trend for higher net CO2 uptake in cut grasslands relative to grazed grasslands is likely because in cut grasslands herbage can be removed with no direct CO2 fluxes, whereas when herbage is grazed, respiration from cows (
55% of animal C intake; Crush et al., 1992; Soussana et al., 2010; Zeeman et al., 2010), and deposited dung (
25–40% of animal intake; Crush et al., 1992; Soussana et al., 2010; Zeeman et al., 2010; Parsons et al., 2013 and references therein) contribute to the measured CO2 losses from the ecosystem. While the difference in pasture use between cutting and grazing can have a large effect on the annual NEP, the differences in C loss pathways are accounted for in NECB calculations (Table 2). Indeed, Soussana et al. (2010) found that sites where the main pasture use was cutting tended to be smaller overall C sinks than sites where the main pasture use was grazing.

18

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

This could be explained by the generally larger off-takes of C through cutting compared to grazing, particularly if manure is not returned at a later date (Soussana et al., 2007, 2010). At Scott Farm, grazing was the main form of C use for 2008–2010, whereas in 2011, substantially more C was permanently exported as silage than during the other years. This likely contributed to the C sink strength in 2011 being the smallest of all four years. 4.3. Scott Farm NEP and NECB compared to other managed temperate grasslands Comparison of annual NEP and NECB for the global dataset (Figs. 4 and 5; Supplementary Table 1) showed that the average NECB (44  33 g C m2 y1) was significantly smaller (p < 0.001, paired t-test) than average NEP (188  44 g C m2 y1), for sites on mineral soils. Lower NECB than NEP in pastoral systems occurs because C exports via non CO2–C pathways (e.g., harvested silage, exported product or DOC leaching) almost always exceed non CO2– C imports via spreading of manure or effluent and import of supplemental feed (Ciais et al., 2010). Indeed, at Scott Farm for three out of the four years, NECB was approximately 150 g C m2 y1 smaller than NEP. However for 2010, when a large amount of effluent was applied to the site, NECB was 27 g C m2 y1 larger than NEP. Only for one other site-year reported in the literature (Chamau, an intensively managed grassland in the Swiss Alps), was NECB found to exceed NEP (Supplementary Table 1), also as a result of large manure applications (Zeeman et al., 2010). The clear difference between NEP and NECB values for most managed temperate grassland sites further emphasised the importance of quantifying non-CO2–C fluxes when calculating the C balance for these systems (Allard et al., 2007; Mudge et al., 2011). Although there were often considerable differences in the size (and sign) of annual NEP and NECB for the site years shown in Fig. 4 and positive NECB's occur across a wide range of NEP’s, we note that NECB generally increased with increasing NEP. This suggests that it may be possible to meet the dual goal of increased pasture production and increased soil C storage in managed temperate grasslands. Aside from site and weather conditions, management practices that could contribute to meeting this dual goal require further identification. For example, when comparing 17 site-years of NECB data from grasslands across Europe, Soussana et al. (2007) found C sink strength to increase with increasing N inputs. Further detailed analysis of site management information and component C fluxes of all sites would be needed to explain the main drivers of the differences in NECB between sites shown in Fig. 4. However, such an analysis would be constrained because insufficient details (e.g., on farm management) are reported in many studies and approaches used for determining non-CO2–C fluxes also differ. For example, some studies ignored components of the NECB such as C exported as milk (e.g., Soussana et al., 2007). Adoption of standard protocols around inclusion of the various non-CO2–C fluxes in NECB calculations would improve comparability between studies and our ability to identify key regulating mechanisms of C storage in grazed pasture systems. 5. Conclusions Despite intra and inter-annual variability in C dynamics at Scott Farm, averaged over the four year study period, the site was a sink for both CO2 (165  51 g C m2 y1), and total C (61  53 g C m2 y1) after non-CO2–C fluxes were accounted for. On annual timescales both NEP and NECB were relatively constant, despite major environmental or management perturbations, such as a severe drought, cultivation and effluent application. This apparent resilience of the CO2 and C balance to perturbations was likely at least partly attributable to the relatively mild temperatures, also in

winter, which allowed pasture growth year-round (in the absence of major disturbances such as drought conditions). Furthermore, weather conditions and management events were inter-related, and there were instances during the four-year study where decisions by the farmer aimed at maintaining steady milk production (for example import of extra feed after pasture renewal; or extra export of pasture as silage in a very productive year) appeared to contribute to maintaining a relatively constant C balance between years as well. The values for NEP and NECB (and the difference between the two) were similar to those found for managed temperate grasslands globally (Fig. 4). Analysis of the global dataset further emphasised the importance of measuring non-CO2–C fluxes, because farms were sinks for CO2 in 78% of site-years, but they were total C sinks after non-CO2–C fluxes were accounted for in only 41% of site-years. In the collated dataset, we also noted a general trend for increased C sink strength with increasing NEP, suggesting that in temperate managed grasslands, it may be possible to achieve increased pasture production (thus milk, meat and fiber production) while sequestering soil C. More studies where full NECB’s are quantified (using consistent protocols for non-CO2–C flux inclusion) should improve our ability to identify the key factors which control longer-term C storage in grazed pasture systems. In particular, side-by-side comparison of treatment and control plots designed to identify management practices aimed at both maintaining (or increasing) production while enhancing soil C storage should be a goal of future research. Further, coupling mechanistic modeling with EC measurements will help constrain flux estimates, and once adequately calibrated, will allow the effect of alternative climatic and management scenarios to be tested. Acknowledgements We would like to thank DairyNZ for providing access to the field site, and DairyNZ staff Errol Thom, John Siemelink, Chris Roach, Carol Leydon-Davis, Chris Glassey, Deanne Waugh, Cameron Clark, and Jim Lancaster for providing production data and information on farm management. We thank Craig Hosking for technical and field assistance, and Isoude Kuijper for careful compilation of grazing, feed and harvest data. Funding for this work was primarily provided by the New Zealand Agricultural Greenhouse Gas Research Centre, with field data collection partially funded by DairyNZ, the University of Waikato and Landcare Research. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.agee.2015.03.011. References Allard, V., Soussana, J.F., Falcimagne, R., Berbigier, P., Bonnefond, J.M., Ceschia, E., D’Hour, P., Henault, C., Laville, P., Martin, C., Pinares-Patino, C., 2007. The role of grazing management for the net biome productivity and greenhouse gas budget (CO2, N2O and CH4) of semi-natural grassland. Agric. Ecosyst. Environ. 121, 47–58. Ammann, C., Spirig, C., Leifeld, J., Neftel, A., 2009. Assessment of the nitrogen and carbon budget of two managed temperate grassland fields. Agric. Ecosyst. Environ. 133, 150–162. Aubinet, M., Moureaux, C., Bodson, B., Dufranne, D., Heinesch, B., Suleau, M., Vancutsem, F., Vilret, A., 2009. Carbon sequestration by a crop over a 4-year sugar beet/winter wheat/seed potato/winter wheat rotation cycle. Agric. For. Meteorol. 149, 407–418. Baker, J.M., Griffis, T.J., 2005. Examining strategies to improve the carbon balance of corn/soybean agriculture using eddy covariance and mass balance techniques. Agric. For. Meteorol. 128, 163–177. Baker, C.J., Richie, W.R., 2007. No-tillage for forage production. In: Baker, C.J., Saxton, K.E. (Eds.), No- Tillage Seeding in Conservation Agriculture. Food and Agriculture Organization of the United Nations and CAB International, Wallingford, UK, pp. 326.

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20 Baldocchi, D.D., 2003. Assessing the eddy covariance technique for evaluating carbon dioxide exchange rates of ecosystems: past, present and future. Global Change Biol. 9, 479–492. Bellamy, P.H., Loveland, P.J., Bradley, R.I., Lark, R.M., Kirk, G.J.D., 2005. Carbon losses from all soils across England and Wales 1978–2003. Nature 437, 245–248. Béziat, P., Ceschia, E., Dedieu, G., 2009. Carbon balance of a three crop succession over two cropland sites in South West France. Agric. For. Meteorol. 149, 1628–1645. Byrne, K.A., Kiely, G., Leahy, P., 2007. Carbon sequestration determined using farm scale carbon balance and eddy covariance. Agric. Ecosyst. Environ. 121, 357–364. Chapin, F., Woodwell, G., Randerson, J., Rastetter, E., Lovett, G., Baldocchi, D., Clark, D., Harmon, M., Schimel, D., Valentini, R., Wirth, C., Aber, J., Cole, J., Goulden, M., Harden, J., Heimann, M., Howarth, R., Matson, P., McGuire, A., Melillo, J., Mooney, H., Neff, J., Houghton, R., Pace, M., Ryan, M., Running, S., Sala, O., Schlesinger, W., Schulze, E.D., 2006. Reconciling carbon-cycle concepts, terminology, and methods. Ecosystems 9, 1041–1050. Ciais, P., Soussana, J.F., Vuichard, N., Luyssaert, S., Don, A., Janssens, I.A., Piao, S.L., Dechow, R., Lathière, J., Maignan, F., Wattenbach, M., Smith, P., Ammann, C., Freibauer, A., Schulze, E.D., the, C.S.T., 2010. The greenhouse gas balance of European grasslands. Biogeosci. Discuss. 7, 5997–6050. Clark, D.A., Caradus, J.R., Monaghan, R.M., Sharp, P., Thorrold, B.S., 2007. Issues and options for future dairy farming in New Zealand. N. Z. J. Agric. Res. 50, 203–221. Conant, R.T., Paustian, K., Elliott, E.T., 2001. Grassland management and conversion into grassland: effects on soil carbon. Ecol. Appl. 11, 343–355. Conant, R.T., Easter, M., Paustian, K., Swan, A., Williams, S., 2007. Impacts of periodic tillage on soil C stocks: a synthesis. Soil Tillage Res. 95, 1–10. Crush, J.R., Waghorn, G.C., Rolston, M.P., 1992. Greenhouse gas emissions from pasture and arable crops grown on a Kairanga soil in the Manawatu, North Island, New Zealand. N. Z. J. Agric. Res. 35, 253–257. Dragoni, D., Schmid, H., Grimmond, C.B., Loescher, H.W., 2007. Uncertainty of annual net ecosystem productivity estimated using eddy covariance flux measurements. J. Geophys. Res. 112. Freibauer, A., Rounsevell, M.D.A., Smith, P., Verhagen, J., 2004. Carbon sequestration in the agricultural soils of Europe. Geoderma 122, 1–23. Gilmanov, T.G., Soussana, J.F., Aires, L., Allard, V., Ammann, C., Balzarolo, M., Barcza, Z., Bernhofer, C., Campbell, C.L., Cernusca, A., Cescatti, A., Clifton-Brown, J., Dirks, B.O.M., Dore, S., Eugster, W., Fuhrer, J., Gimeno, C., Gruenwald, T., Haszpra, L., Hensen, A., Ibrom, A., Jacobs, A.F.G., Jones, M.B., Lanigan, G., Laurila, T., Lohila, A., Manca, G., Marcolla, B., Nagy, Z., Pilegaard, K., Pinter, K., Pio, C., Raschi, A., Rogiers, N., Sanz, M.J., Stefani, P., Sutton, M., Tuba, Z., Valentini, R., Williams, M. L., Wohlfahrt, G., 2007. Partitioning European grassland net ecosystem CO2 exchange into gross primary productivity and ecosystem respiration using light response function analysis. Agric. Ecosyst. Environ. 121, 93–120. Göckede, M., Foken, T., Aubinet, M., Aurela, M., Banza, J., Bernhofer, C., Bonnefond, J. M., Brunet, Y., Carrara, A., Clement, R., Dellwik, E., Elbers, J., Eugster, W., Fuhrer, J., Granier, A., Grünwald, T., Heinesch, B., Janssens, I.A., Knohl, A., Koeble, R., Laurila, T., Longdoz, B., Manca, G., Marek, M., Markkanen, T., Mateus, J., Matteucci, G., Mauder, M., Migliavacca, M., Minerbi, S., Moncrieff, J., Montagnani, L., Moors, E., Ourcival, J.M., Papale, D., Pereira, J., Pilegaard, K., Pita, G., Rambal, S., Rebmann, C., Rodrigues, A., Rotenberg, E., Sanz, M.J., Sedlak, P., Seufert, G., Siebicke, L., Soussana, J.F., Valentini, R., Vesala, T., Verbeeck, H., Yakir, D., 2008. Quality control of CarboEurope flux data – Part 1: coupling footprint analyses with flux data quality assessment to evaluate sites in forest ecosystems. Biogeosciences 5, 433–450. Govaerts, B., Verhulst, N., Castellanos-Navarrete, A., Sayre, K.D., Dixon, J., Dendooven, L., 2009. Conservation agriculture and soil carbon sequestration: between myth and farmer reality. Crit. Rev. Plant Sci. 28, 97–122. Hewitt, A.E., 1993. New Zealand soil classification. Landcare Science Series No. 1. Manaaki-Whenua Press, Lincoln, New Zealand. Hollinger, D.Y., Richardson, A.D., 2005. Uncertainty in eddy covariance measurements and its application to physiological models. Tree Physiol. 25, 873–885. Hussain, M.Z., Grünwald, T., Tenhunen, J.D., Li, Y.L., Mirzae, H., Bernhofer, C., Otieno, D., Dinh, N.Q., Schmidt, M., Wartinger, M., Owen, K., 2011. Summer drought influence on CO2 and water fluxes of extensively managed grassland in Germany. Agric. Ecosyst. Environ. 141, 67–76. Jaksic, V., Kiely, G., Albertson, J., Oren, R., Katul, G., Leahy, P., Byrne, K.A., 2006. Net ecosystem exchange of grassland in contrasting wet and dry years. Agric. For. Meteorol. 139, 323–334. Jones, M.B., 2010. Potential for carbon sequestration in temperate grassland soils. In: Abberton, M., Conant, R., Batello, C. (Eds.), Grassland Carbon Sequestration: Management, Policy and Economics – Proceedings of the Workshop on the Role of Grassland Carbon Sequestration in the Mitigation of Climate Change. Plant Production and Protection Division Food and Agriculture Organization of the United Nations (FAO), Rome. Kirschbaum, M.U.F., Rutledge, S., Kuijper, I.A., Mudge, P.L., Puche, N., Wall, A.M., Roach, C.G., Schipper, L.A., Campbell, D.I., 2015. Modelling carbon and water exchange of a grazed pasture in New Zealand constrained by eddy covariance measurements. Sci. Total Environ. 512–513, 273–286. Kormann, R., Meixner, F.X., 2001. An analytical footprint model for non-neutral stratification. Bound. Layer Meteorol. 99, 207–224. Lal, R., 2004a. Agricultural activities and the global carbon cycle. Nutr. Cycling Agroecosyst. 70, 103–116. Lal, R., 2004b. Soil carbon sequestration to mitigate climate change. Geoderma 123, 1–22.

19

MacLeod, C.J., Moller, H., 2006. Intensification and diversification of New Zealand agriculture since 1960: an evaluation of current indicators of land use change. Agric. Ecosyst. Environ. 115, 201–218. McMillen, R.T., 1986. A BASIC Program for EC in non-simple terrain. NOAA Tech Memo ERL ARL-147. Meersmans, J., Van Wesemael, B., De Ridder, F., Fallas Dotti, M., De Baets, S., Van Molle, M., 2009. Changes in organic carbon distribution with depth in agricultural soils in northern Belgium, 1960–2006. Global Change Biol. 15, 2739–2750. Moore, C.J., 1986. Frequency response corrections for eddy correlation systems. Bound. Layer Meteorol. 37, 17–35. Mudge, P.L., Wallace, D.F., Rutledge, S., Campbell, D.I., Schipper, L.A., Hosking, C.L., 2011. Carbon balance of an intensively grazed temperate pasture in two climatically contrasting years. Agric. Ecosyst. Environ. 144, 271–280. Parsons, A., Rowarth, J., Thornley, J., Newton, P., 2011. Primary production of grasslands, herbage accumulation and use, and impacts of climate change. In: Lemaire, G., Hodgson, J., Chabbi, A. (Eds.), Grassland Productivity and Ecosystem Services Wallingford, UK: CAB International, p. 3–18. Parsons, A.J., Thornley, J.H.M., Newton, P.C.D., Rasmussen, S., Rowarth, J.S., 2013. Soil carbon dynamics: the effects of nitrogen input, intake demand and off-take by animals. Sci. Total Environ. 465, 205–215. Prescher, A.-K., Grünwald, T., Bernhofer, C., 2010. Land use regulates carbon budgets in eastern Germany: from NEE to NBP. Agric. For. Meteorol. 150, 1016–1025. Rees, R.M., Bingham, I.J., Baddeley, J.A., Watson, C.A., 2005. The role of plants and land management in sequestering soil carbon in temperate arable and grassland ecosystems. Geoderma 128, 130–154. Reichstein, M., Falge, E., Baldocchi, D., Papale, D., Aubinet, M., Berbigier, P., Bernhofer, C., Buchmann, N., Gilmanov, T., Granier, A., Grunwald, T., Havrankova, K., Ilvesniemi, H., Janous, D., Knohl, A., Laurila, T., Lohila, A., Loustau, D., Matteucci, G., Meyers, T., Miglietta, F., Ourcival, J.M., Pumpanen, J., Rambal, S., Rotenberg, E., Sanz, M., Tenhunen, J., Seufert, G., Vaccari, F., Vesala, T., Yakir, D., Valentini, R., 2005. On the separation of net ecosystem exchange into assimilation and ecosystem respiration: review and improved algorithm. Global Change Biol. 11, 1424–1439. Reicosky, D.C., Lindstrom, M.J., 1993. Fall tillage method: effect on short-term carbon dioxide flux from soil. Agron. J. 85, 1237–1243. Rogiers, N., Eugster, W., Furger, M., Siegwolf, R., 2005. Effect of land management on ecosystem carbon fluxes at a subalpine grassland site in the Swiss Alps. Theor. Appl. Climatol. 80, 187–203. Rutledge, S., Mudge, P.L., Wallace, D.F., Campbell, D.I., Woodward, S.L., Wall, A.M., Schipper, L.A., 2014. CO2 emissions following cultivation of a temperate permanent pasture. Agric. Ecosyst. Environ. 184, 21–33. Schipper, L.A., Parfitt, R.L., Fraser, S., Littler, R.A., Baisden, W.T., Ross, C., 2014. Soil order and grazing management effects on changes in soil C and N in New Zealand pastures. Agric. Ecosyst. Environ. 184, 67–75. Schnabel, R.R., Franzluebbers, A.J., Stout, W.L., Sanderson, M.A., Stuedemann, J.A., 2001. The effect of pasture management practices. In: Follett, R.F., Kimble, J.M., Lal, R. (Eds.), The Potential of U.S. Grazing Lands to Sequester Carbon and Mitigate the Greenhouse Effect. Lewis Publishers/CRC Press, Boca Raton – London – New York – Washington. Schotanus, P., Nieuwstadt, F.T.M., de Bruin, H.A.R., 1983. Temperature measurement with a sonic anemometer and its application to heat and moisture fluxes. Bound. Layer Meteorol. 26, 81–90. Skinner, R.H., 2008. High biomass removal limits carbon sequestration potential of mature temperate pastures. J. Environ. Qual. 37, 1319–1326. Smith, P., 2004. How long before a change in soil organic carbon can be detected? Global Change Biol. 10, 1878. Smith, P., Martino, D., Cai, Z., Gwary, D., Janzen, H., Kumar, P., McCarl, B., Ogle, S., O'Mara, F., Rice, C., Scholes, B., Sirotenko, O., Howden, M., McAllister, T., Pan, G., Romanenkov, V., Schneider, U., Towprayoon, S., Wattenbach, M., Smith, J., 2008. Greenhouse gas mitigation in agriculture. Phil. Trans. R. Soc. B: Biol. Sci. 363, 789–813. Smith, P., Lanigan, G., Kutsch, W.L., Buchmann, N., Eugster, W., Aubinet, M., Ceschia, E., Béziat, P., Yeluripati, J.B., Osborne, B., Moors, E.J., Brut, A., Wattenbach, M., Saunders, M., Jones, M., 2010. Measurements necessary for assessing the net ecosystem carbon budget of croplands. Agric. Ecosyst. Environ. 139, 302–315. Soussana, J.F., Loiseau, P., Vuichard, N., Ceschia, E., Balesdent, J., Chevallier, T., Arrouays, D., 2004. Carbon cycling and sequestration opportunities in temperate grasslands. Soil Use Manage. 20, 219–230. Soussana, J.F., Allard, V., Pilegaard, K., Ambus, P., Amman, C., Campbell, C., Ceschia, E., Clifton-Brown, J., Czobel, S., Domingues, R., Flechard, C., Fuhrer, J., Hensen, A., Horvath, L., Jones, M., Kasper, G., Martin, C., Nagy, Z., Neftel, A., Raschi, A., Baronti, S., Rees, R.M., Skiba, U., Stefani, P., Manca, G., Sutton, M., Tuba, Z., Valentini, R., 2007. Full accounting of the greenhouse gas (CO2, N2O, CH4) budget of nine European grassland sites. Agric. Ecosyst. Environ. 121, 121–134. Soussana, J.F., Tallec, T., Blandfort, V., 2010. Mitigating the greenhouse gas balance of ruminant production systems through carbon sequestration in grasslands. Animal 4, 334–350. van Wesemael, B., Paustian, K., Meersmans, J., Goidts, E., Barancikova, G., Easter, M., 2010. Agricultural management explains historic changes in regional soil carbon stocks. Proc. Natl. Acad. Sci. 107, 14926–14930. Veenendaal, E.M., Kolle, O., Leffelaar, P.A., Schrier-Uijl, A.P., Van Huissteden, J., Van Walsem, J., Möller, F., Berendse, F., 2007. CO2 exchange and carbon balance in two grassland sites on eutrophic drained peat soils. Biogeosciences 4, 1027–1040.

20

S. Rutledge et al. / Agriculture, Ecosystems and Environment 206 (2015) 10–20

Watson, R.T., Noble, I.R., Bolin, B., Ravindranath, N.H., Verardo, D.J., Dokken, D.J., 2000. Land Use, Land-Use Change and Forestry. Cambridge University Press, Cambridge. Webb, E.K., Pearman, G.J., Leuning, R., 1980. Correction of flux measurements for density effects due to heat and water vapour transfer. Q. J. R. Meteorol. Soc. 106, 85–100. Willems, A.B., Augustenborg, C.A., Hepp, S., Lanigan, G., Hochstrasser, T., Kammann, C., Müller, C., 2011. Carbon dioxide emissions from spring ploughing of grassland in Ireland. Agric. Ecosyst. Environ. 144, 347–351. Wohlfahrt, G., Hammerle, A., Haslwanter, A., Bahn, M., Tappeiner, U., Cernusca, A., 2008. Seasonal and inter-annual variability of the net ecosystem CO2 exchange of a temperate mountain grassland: effects of weather and management. J. Geophys. Res. 11, 3. doi:http://dx.doi.org/10.1029/2007JD009286. Yi, C., Ricciuto, D., Li, R., Wolbeck, J., Xu, X., Nilsson, M., Aires, L., Albertson, J.D., Ammann, C., Arain, M.A., Araujo d, A.C., Aubinet, M., Aurela, M., Barcza, Z., Barr, A., Berbigier, P., Beringer, J., Bernhofer, C., Black, A.T., Bolstad, P.V., Bosveld, F.C., Broadmeadow, M.S.J., Buchmann, N., Burns, S.P., Cellier, P., Chen, J., Chen, J., Ciais, P., Clement, R., Cook, B.D., Curtis, P.S., Dail, D.B., Dellwik, E., Delpierre, N., Desai, A.R., Dore, S., Dragoni, D., Drake, B.G., Dufrêne, E., Dunn, A., Elbers, J., Eugster, W., Falk, M., Feigenwinter, C., Flanagan, L.B., Foken, T., Frank, J., Fuhrer, J., Gianelle, D., Goldstein, A., Goulden, M., Granier, A., Grünwald, T., Gu, L., Guo, H.,

Hammerle, A., Han, S., Hanan, N.P., Haszpra, L., Heinesch, B., Helfter, C., Hendriks, D., Hutley, L.B., Ibrom, A., Jacobs, C., Johansson, T., Jongen, M., Gabrie, K., Kiely, G., Klumpp, K., Knohl, A., Kolb, T., Kutsch, W.L., Lafleur, P., Laurila, T., Leuning, R., Lindroth, A., Liu, H., Loubet, B., Manca, G., Marek, M., Margolis, H.A., Martin, T.A., Massman, W.J., Matamala, R., Matteucci, G., McCaughey, H., Merbold, L., Meyers, T., Migliavacca, M., Miglietta, F., Misson, L., Mölder, M., Moncrieff, J., Monson, R. K., Montagnani, L., Montes-Helu, M., Moors, E., Moureaux, C., Mukelabai, M.M., Munger, J.W., Myklebust, M., Nagy, Z., Noormets, A., Oechel, W., Oren, R., Pallardy, S.G., Paw U, K.T., Pereira, J.S., Pilegaard, K., Pintér, K., Pio, C., Pita, G., Powell, T.L., Rambal, S., Randerson, J.T., von Randow, C., Rebmann, C., Rinne, J., Rossi, F., Roulet, N., Ryel, R.J., Sagerfors, J., Saigusa, N., Sanz, M.J., Mugnozza, G.-S., Schmid, H.P., Seufert, G., Siqueira, M., Soussana, J.-F., Starr, G., Sutton, M.A., Tenhunen, J., Tuba, Z., Tuovinen, J.-P., Valentini, R., Vogel, C.S., Wang, J., Wang, S., Wang, W., Welp, L.R., Wen, X., Wharton, S., Wilkinson, M., Williams, C.A., Wohlfahrt, G., Yamamoto, S., Yu, G., Zampedri, R., Zhao, B., Zhao, X., 2010. Climate control of terrestrial carbon exchange across biomes and continents. Environ. Res. Lett. 5, 34007. Zeeman, M.J., Hiller, R., Gilgen, A.K., Michna, P., Plüss, P., Buchmann, N., Eugster, W., 2010. Management and climate impacts on net CO2 fluxes and carbon budgets of three grasslands along an elevational gradient in Switzerland. Agric. For. Meteorol. 150, 519–530.