Soil greenhouse gas fluxes and global warming potential in four high-yielding maize systems

Global Change Biology (2007) 13, 1972–1988, doi: 10.1111/j.1365-2486.2007.01421.x Soil greenhouse gas fluxes and global warming potential in four high-yielding maize systems M . A . A . A D V I E N T O - B O R B E , M . L . H A D D I X , D . L . B I N D E R , D . T. WA LT E R S and A . D O B E R M A N N 1 Department of Agronomy and Horticulture, University of Nebraska, Lincoln, NE 68586-0915, USA Abstract Crop intensification is often thought to increase greenhouse gas (GHG) emissions, but studies in which crop management is optimized to exploit crop yield potential are rare. We conducted a field study in eastern Nebraska, USA to quantify GHG emissions, changes in soil organic carbon (SOC) and the net global warming potential (GWP) in four irrigated systems: continuous maize with recommended best management practices (CCrec) or intensive management (CC-int) and maize–soybean rotation with recommended (CS-rec) or intensive management (CS-int). Grain yields of maize and soybean were generally within 80–100% of the estimated site yield potential. Large soil surface carbon dioxide (CO2) fluxes were mostly associated with rapid crop growth, high temperature and high soil water content. Within each crop rotation, soil CO2 efflux under intensive management was not consistently higher than with recommended management. Owing to differences in residue inputs, SOC increased in the two continuous maize systems, but decreased in CS-rec or remained unchanged in CS-int. N2O emission peaks were mainly associated with high temperature and high soil water content resulting from rainfall or irrigation events, but less clearly related to soil NO3-N levels. N2O fluxes in intensively managed systems were only occasionally greater than those measured in the CC-rec and CS-rec systems. Fertilizer-induced N2O emissions ranged from 1.9% to 3.5% in 2003, from 0.8% to 1.5% in 2004 and from 0.4% to 0.5% in 2005, with no consistent differences among the four systems. All four cropping systems where net sources of GHG. However, due to increased soil C sequestration continuous maize systems had lower GWP than maize– soybean systems and intensive management did not cause a significant increase in GWP. Converting maize grain to ethanol in the two continuous maize systems resulted in a net reduction in life cycle GHG emissions of maize ethanol relative to petrol-based gasoline by 33–38%. Our study provided evidence that net GHG emissions from agricultural systems can be kept low when management is optimized toward better exploitation of the yield potential. Major components for this included (i) choosing the right combination of adopted varieties, planting date and plant population to maximize crop biomass productivity, (ii) tactical water and nitrogen (N) management decisions that contributed to high N use efficiency and avoided extreme N2O emissions, and (iii) a deep tillage and residue management approach that favored the build-up of soil organic matter from large amounts of crop residues returned. Keywords: agriculture, C sequestration, global warming potential, intensive cropping systems, maize, nitrous oxide Received 22 June 2006; revised version received 1 May 2007 and accepted 25 May 2007 Introduction Correspondence: Achim Dobermann, tel. 11 402 472 1501, fax 11 402 472 7904, e-mail: a.dobermann@cgiar.org 1 Present address: Achim Dobermann, International Rice Research Institute (IRRI), DAPO Box 7777, Manila 1271, The Philippines. 1972 Rainfed and irrigated continuous maize (Zea mays L.) or maize–soybean (Glycine max. [L] Merr) systems in the North American Corn Belt account for more than onethird of the world’s maize and soybean production r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd S O I L G H G F L U X E S A N D G L O B A L WA R M I N G (Dobermann & Cassman, 2002). Meeting the future global demand for maize and soybean, including the rapidly rising feedstock demand for biofuel production, will largely have to be achieved through yield increases (Cassman et al., 2003). Average crop yields may have to approach 80% of the yield potential or more, particularly in areas with favorable rainfall or irrigation. At issue is whether high crop yields can be achieved without increasing greenhouse gas (GHG) emissions from agricultural land. A central hypothesis for such an ecological intensification of agriculture is that an optimal balance of high productivity, sustainability and minimal environmental impact can be achieved by optimizing management toward better exploitation of crop yield potential. In agricultural systems under temperate climate, net emissions of carbon dioxide (CO2) and nitrous oxide (N2O) from the soil surface are of particular interest because they are the major components of the net global warming potential (GWP) of cultivated land (Robertson & Grace, 2004). Significant potential may exist for managing CO2 and N2O fluxes through shifts in land use or cropping systems as well as by improving crop, soil and fertilizer management practices. The latter may range widely, from measures to increase crop yield to better management of critical inputs such as water and nitrogen (N) or conservation agriculture methods for increasing soil C sequestration. To assess such options requires a full-cost accounting of the GWP of agricultural systems, including net changes in soil organic carbon (SOC), intrinsic C costs associated with crop production, and net emissions of N2O and CH4 (Robertson et al., 2000; Robertson & Grace, 2004). Little information is available for cropping systems that are designed to explore the upper limits of agriculture. A long-term experiment was established in 1999 in eastern Nebraska, USA to evaluate the agroecological performance of maize-based cropping systems with different levels of management intensity, aiming at crop yields within 80–95% of the current climatic-genetic yield potential. In this paper, we report on (i) GHG fluxes, (ii) changes in SOC and (iii) the net GWP in four high-yielding maize-based cropping systems. Materials and methods Field experiment This study was conducted in four treatments of the Ecological Intensification of Irrigated Maize-based Cropping Systems Experiment at the University of Nebraska–Lincoln East Campus in Lincoln, NE (Latitude 40.82; Longitude 96.65; Elevation 357 m). The experiment was established in 1999 to evaluate 1973 resource-efficient management concepts for achieving crop yields that approach the climatic yield potential. The soil is a deep Kennebec silty clay loam (fine-silty, mixed, superactive, mesic Cumulic Hapludolls) of generally high quality. Average soil properties measured in March 2003 in the 0–0.3 m layer were pHwater 6.14, 370 g kg 1 clay, 580 g kg 1 silt, 50 g kg 1 sand, 18.2 g kg 1 organic C, 27.4 cmolc kg 1 CEC, 0.77 cmolc kg 1 exchangeable K, and 73 mg kg 1 available Bray-1 P. The field experiment was conducted in a split–split plot randomized complete block design (four replicates) with three factors: three crop rotations as main plots (CC, continuous maize; CS, maize–soybean with maize in even years; SC, maize–soybean with maize in odd years), three plant population densities as subplots and two levels of nutrient management as sub-subplots. Four management systems were selected for detailed gas flux measurements from 2003 to 2005: (1) CC-rec, continuous maize with recommended management, (2) CC-int, continuous maize with intensified management; (3) CS-rec, maize–soybean rotation with recommended management; and (4) CS-int, maize–soybean rotation with intensified management. Management practices are summarized in Table 1. The CC-rec and CS-rec systems represent recommended plant populations, nutrient and water management practices for growing irrigated maize and soybean in eastern Nebraska, aiming at maize yields of about 14 Mg ha 1. In the CC-int and CS-int systems, management aimed at achieving maize yields of 18 Mg ha 1, which is equivalent to the climatic yield potential at this site in favorable years. Key additional measures for that included increased plant populations, increased fertilizer rates, and more frequent N applications to achieve high N use efficiency at high input levels (Table 1). In all systems studied, annual amounts of fertilizer-N (Table 2) were calculated using an algorithm which includes yield goal, soil organic matter content, residual NO3-N in spring, and credits given to legumes as previous crops, manure, or N applied with irrigation water (Shapiro et al., 2003). Residual soil NO3-N in 0–1.2 m depth was measured in April of each year in each plot and was the main factor for annual adjustment of N rates. In CC-rec and CS-rec, N fertilizer was applied preplant (50–60%) and at six-leaf stage of maize, both as ammonium nitrate and incorporated to a depth of 5–10 cm by field cultivator. In CC-int and CSint, N application to maize was done in four split applications (preplant, V6, V10, shortly before tasseling). The two late doses were hand-applied between rows followed by sprinkler irrigation. Since 2001 (CCint) and 2004 (CS-int), N management in the two intensive systems included an additional application of 50 kg N ha 1 sprayed as urea–ammonium-nitrate r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 CC-rec* CC-int* CS-rec* CS-int* r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 Harvest and crop residues Tillage Lime and micronutrients Irrigation Weed management Pest management Varieties Planting dates Plant population, plants (m 2) Row spacing (m) Combine for removal of grain; crop residues incorporated with fall plowing Fall plowing with a conservation tillage plow, 0.25–0.3 m deep, 20% residue coverage Blanket applied every 5–6 years (initial pH correction 1999–2001: 9 Mg lime ha 1) Sprinkler. M, 300–450 mm in 2003–2005 (11–14 doses); S, 180 mm in 2005 (four doses) Pre-emergence herbicide, field cultivation 4 weeks after emergence, hand weeding Bt-hybrids (M), insecticide and fungicide treatment of seed, insecticide at planting Maize: Pioneer 31N28 (2003–2004), Pioneer 31G68; Soybean: NE3001 Maize: May 10–15; Soybean: May 1–5 7–9 9–11 7–9(M); 25–28(S) 0.76 0.76 0.76 Annual nitrogen (N) application (kg N ha 1) N applications during growing season N application on maize residue in fall Annual P application (kg P ha 1) Annual K application (kg K ha 1) Long-term averages, 2000–2005 Annual fertilizer N input ( N ha 1) Annual total crop N uptake (kg N ha 1) Annual N removal with grain (kg N ha 1) Grain yield of maize (Mg ha 1) Grain yield of soybean (Mg ha 1) 180–240 250–310 130–140(M); 0(S) 9–11(M); 28–35(S) 0.76(M); 0.38 or 0.76(S) 230–250(M); 80–130(S) 2 4 2(M); 0(S) 4(M); 1(S) None Since 2001 None Since 2004 0 0 45 85 0 0 45 85 201 299 70 172 260 317 324 346 176 194 229 238 13.95 – 14.95 – 14.71 4.89 15.61 5.02 *M and S indicate maize and soybean crops, respectively. 1974 M . A . A . A D V I E N T O - B O R B E et al. Table 1 Crop management practices, grain yields, and crop residue input in continuous maize (CC) and maize-soybean (CS) systems with recommended (-rec) or intensive management (-int) S O I L G H G F L U X E S A N D G L O B A L WA R M I N G 1975 Table 2 Residual soil NO3-N levels before planting, rates of nitrogen application and crop yields during 2003–2005 in four cropping systems of the ecological intensification experiment at Lincoln, NE Continuous maize Year Previous crop Soil NO3 N before planting (kg N ha 1)* Crop Fertilizer N (kg N ha 1) Grain yield (Mg ha 1) CC-rec 2003 2004 2005 Maize Maize Maize 2003 2004 2005 2003 2004 2005 Maize Maize Maize Maize Maize Maize 75 29 16 180 200 240 16.01 15.52 13.85 Maize-soybean CC-int CS-rec 172 164 56 Soybean Soybean Maize 250 280 310 15.83 16.74 11.99 Maize Maize Soybean Maize Maize Soybean 67 26 31 130 140 0 16.80 16.35 5.07 CS-int 75 53 35 250 230 130 17.53 18.01 5.31 *Residual soil NO3-N in 0 to 1.2 m depth, measured in April, before N application and planting. solution on the maize residue before plowing in fall. This was done to facilitate better decomposition and humification of maize residue. N rates of the succeeding crop were adjusted accordingly, based on residual soil NO3 measured in spring. Under soybean, N fertilizer was only applied in the CS-int system at R3.5 stage. Owing to high soil test levels, no P and K fertilizer was applied in CC-rec and CS-rec, but maize and soybean grown in CC-int and CS-int received annual applications of P and K to replenish crop removal (Table 1). In 2005, gas measurements were also conducted in an adjacent grassland strip (50 m  10 m). The grassland was originally a mowed field divider and was primarily dominated by Indian grass (Sorghastrum nutans) and switch grass (Panicum virgatum). In 2000, the grassland was converted to a tallgrass prairie by seeding native tallgrass species such as big blue stem (Andropogon gerardii), little bluestem (Schizachyrium scoparium) and side oats grama (Bouteloua curtipendula). Measurements CO2, N2O and CH4 fluxes from the soil surface were measured weekly to biweekly during the growing season and less frequently in the nongrowing season using static flux chambers and a model 1312 Infrared Photoacoustic Spectroscopy (PAS) gas analyzer (Innova Air Tech Instruments, Ballerup, Denmark). Adjustments in sampling dates and frequency were made to include specific events such as heavy rainfall, irrigation, high temperature, N application or tillage. Fluxes were measured on 64 occasions: CC-rec and CC-int systems, throughout all three maize crops grown from 2003 to 2005; CS-rec and CS-int systems, for maize following soybean (2003 and 2004) and for soybean following maize (2005). Flux measurements were always conducted between 08:00 to 13:00 hours to minimize diurnal variation in the flux pattern. On each sampling day, the sequence of gas measurements in the 16 plots (four cropping systems  four replicates) was randomized to avoid bias due to rising temperature during the morning hours. Chambers were rectangular steel frames with dimensions of 0.75 m  0.1 m  0.2 m, insulated with Styrofoams (Dow Chemical Co., Midland, MI, USA) and placed across crop rows. The chambers were pushed into the soil to a depth of 0.1 m so that the headspace height remaining above the soil surface was 0.1 m. For the grassland site, three smaller chambers (0.4 m  0.1 m  0.2 m) were placed at the middle of the grassland area at least 6 m away from each other. During measurements, a vented lid with three sampling ports was placed on a chamber, sealed, and connected in a closed-loop system with the PAS gas analyzer. Total measurement time was 14 min with 2 min sampling intervals and a sampling rate of 1.8 L min 1. Fluxes of CO2, N2O and CH4 were computed by fitting a linear regression of gas concentration against time after chamber closure. The time period used for flux rate calculation was the 6–14 min time segment (five data points). Simultaneously with each flux measurement, soil temperature and volumetric soil water content were measured at mid-points of 0–0.05, 0.05–0.15 and 0.15– 0.30 m soil depths using a digital thermometer (Spectrum Technologies Inc., IL, USA) and a MLX-2 Theta probe moisture sensor (Dynamax Inc., Houston, TX, USA), respectively. In 2003, soil electrical conductivity (EC) was also measured in situ using a pencil conductivity probe meter (Hanna Instruments Dist WP, Woonsocket, RI, USA). Soil samples were collected during r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 1976 M . A . A . A D V I E N T O - B O R B E et al. selected flux measurement events to determine soil water content, bulk density, pH, EC and NO3-N dynamics in 0–0.05, 0.05–0.15 and 0.15–0.30 m depths during 2003 and 2005 (only NO3-N, density and water content). Total porosity and water-filled pore space (WFPS) were calculated based on water content, soil bulk density and a particle density of 2.65 g cm 3. Changes in SOC and total soil N (TSN) in the top 0.3 m were measured by collecting soil samples in June of 2000 and 2005. Within each 15 m  6 m plot eight 32 mm  300 mm soil cores were collected, split into 0–0.15 and 0.15–0.30 m segments and combined into one composite sample per depth and plot. All soil samples were air-dried, completely passed through a 2 mm sieve and recognizable undecomposed organic matter particles were removed. A subsample was fineground to 100 mesh for SOC and TSN analysis by dry combustion (ECS 4010, Costech Analytical Technologies Inc., Valencia, CA, USA). Laboratory precision was 0.6 g C kg 1 soil. Bulk density was measured from two separate cores per plot. Total SOC and TSN stocks (g m 2) were calculated for an equivalent dry soil mass of 400 kg m 2 (approximately 0–0.3 m depth) as described by Gifford & Roderick (2003). Grain yields of maize (reported at 15.5% grain moisture content) and soybean (13% seed moisture content) as well as crop N and C uptake in different plant parts were determined for a harvest area of two rows 9 m in each plot. Data analysis Seasonal and annual GHG fluxes were calculated by summing daily fluxes over time. Empirical equations describing the relationship between measured fluxes and environmental variables were used to predict daily GHG fluxes for days between measurements. Measured flux data were used to fit separate prediction models for each treatment and year. Four models were used for predicting CO2 emissions on nonmeasurement days: (i) exponential equation based on hourly air temperature (Fahey et al., 2005), (ii) exponential equation based on hourly soil temperature in 0.1 m depth, (iii) the empirical model with soil temperature proposed by Raich & Mora (2005) and (iv) an equation including volumetric water content, soil temperature and leaf area index (LAI) proposed by Norman et al. (1992). Hourly air and soil temperature (0.1 m) readings were obtained from an automatic weather station located 250 m away from the study field. Daily LAI values needed for estimating CO2 fluxes with the Norman equation were obtained from simulating maize growth with the Hybrid-Maize simulation model (Yang et al., 2004) or by measurements of LAI in soybean during 2005 (Setiyono et al., 2007). Because none of the different prediction equations showed a consistent advantage across treatments and years (R2 values varied from 0.6 to 0.75 with no apparent prediction bias), the estimates of CO2 flux obtained for each treatment-year with the four different empirical equations were averaged. Treatment-specific seasonal and annual estimates of soil N2O emissions were based on averages of two exponential equations for predicting N2O fluxes on nonmeasurement days with either hourly air or soil temperature (0.1 m) as external driving variable (Flessa et al., 2002). Partial correlation analysis showed that temperature was the main driving variable for N2O emissions, whereas inclusion of soil water content or soil N content did not improve the prediction models on a treatment  year basis. In other words, fitting of treatment-yearspecific temperature-driven models accounted for treatment differences in soil moisture and soil nitrate. The R2 values of the different prediction models used to predict the measured N2O fluxes ranged from 0.5 to 0.55. Methane fluxes were generally small and integrated over time by linear interpolation across measurement events because no significant relationships could be established with measured environmental factors. To evaluate the effect of management systems on GHG fluxes during the growing season, SAS programs for split–split unbalanced tests were used (SAS Institute Inc., 1999, SAS System Version 8.0, Cary, NC, USA). Main effects due to crop rotation, as well as interaction and random effects of fertility level, plant density, crop rotation and blocking were analyzed using the PROC MIXED covariance test. Relationships of soil and environmental variables with N2O and CO2 emissions were assessed using partial correlation analysis. Oneway analysis of variance was performed to assess treatment differences in crop residue inputs and changes in SOC and TSN over time. Because of the large spatial variability in SOC and TSN, samples collected from two plant population subplots were pooled for this analysis, (i.e. – rec systems included plots with recommended fertilizer input and low or medium populations), whereas – int systems included plots with high fertilizer input and medium or high populations. Within each of the four systems, there was no significant difference in crop productivity, residue input, or SOC between the two population levels pooled. Calculation of net GWP followed Robertson & Grace (2004) and included intrinsic C costs associated with all production inputs and field operations (averages for maize and soybean crops grown during 2000–2005), measured changes in SOC (2000–2005), and measured fluxes of N2O and CH4 (2003–2005). We used the conversion factors in the ‘ethanol today’ scenario of the ERG Biofuel Analysis Meta-Model (EBAMM) (Farrell r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 S O I L G H G F L U X E S A N D G L O B A L WA R M I N G et al., 2006) for calculating CO2-equivalents of emissions associated with fertilizers, lime, transport, pesticides, herbicides, seed, fuel, electricity, irrigation and farm machinery. Modifications made in EBAMM included use of actual production inputs, measured soil CH4 emission or uptake, measured net sequestration of atmospheric CO2 in SOC, and measured N2O efflux instead of a standard factor for fertilizer-induced N2O emission. For the GWP calculation, we assumed a longterm annual liming rate of 0.5 Mg ha 1 for CC-rec and CS-rec and 0.75 Mg ha 1 for CC-int and CS-int systems. The 0.5 Mg ha 1 value for the two recommended management systems is close to the current national average lime application rate. Higher N rates in the intensive systems require increased liming to maintain soil pH in an optimal range. GWP of soil surface N2O emissions and CH4 oxidation was based on the measured annual fluxes: CC-rec and CC-int: average for maize grown from 2003 to 2005; CS-rec and CS-int – average for maize grown from 2003 and 2004 and soybean grown in 2005. A 100year time horizon was assumed, (i.e. radiative forcing potential relative to CO2 was 296 for N2O and 23 for CH4; IPCC, 2001). In a second calculation we used EBAMM to assess the net reduction in GHG emissions for a scenario in which all maize grain produced in the two continuous maize systems would be used for ethanol production, assuming a biorefinery yield of 0.399 L ethanol kg 1 maize grain and accounting for the total energy input in the biorefinery phase, as well as ethanol coproduct credits (Farrell et al., 2006). Maize yields used for this calculation were averages of 2000–2005 (Table 1). Results Climate and crop yields Average crop yields in this long-term experiment (Table 1) were close to the yield potential of soybean and maize at this location and significantly higher than national or state averages. Large yield fluctuations occurred during the 2003–2005 period due to variation in weather. Annual average temperatures in 2003 and 2004 (11.3–11.5 1C) were cooler than in 2005 (12.3 1C). Temperature differences occurred during the main growing season (May 20 to September 20), with temperatures averaging 22.5 1C in 2003, 21.8 1C in 2004 and 24.4 1C in 2005. Total rainfall was similar in all 3 years, ranging from 552 to 640 mm for the whole year or 285 to 327 mm during the growing season. A total of 71–87 rainfall events occurred per year, with about half of those during the growing season. Assuming no limitations due to water, nutrients, pests or other yieldreducing factors, maize yield potential predicted with the Hybrid-Maize model (Yang et al., 2004) for the 1977 intensive treatments with high plant population was 18.0 Mg ha 1 in 2003, 18.8 Mg ha 1 in 2004 and 14.2 Mg ha 1 in 2005. Actual maize yields (Table 2) in CC-int and CS-int were within 84–97% of the simulated yield potential in 2003 and 2004. In CC-rec and CS-rec, maize yields in 2003 and 2004 (15.5–16.8 Mg ha 1) were within 83–93% of the simulated yield potential in 2003 and 2004 or 97% in 2005. Lower maize yields in 2005 were caused by heat stress during silking (affecting kernel set of maize) and due to a shorter grain filling period. Owing to the high plant population density, the CC-int treatment was more negatively affected by these heat stresses than the CC-rec system. Maize following soybean (CS) yielded about 5–11% higher than continuous maize (CC), primarily due to fewer problems with crop establishment and some insect pests. Soybean yields (Table 1) averaged about 5 Mg ha 1, with maximum yield of 5.9 Mg ha 1 measured in 2001. Soybean yield potential under the conditions of eastern Nebraska is about 6–7 Mg ha 1 (Specht et al., 1999). N use efficiency in maize grown in 2003–2004, calculated as amount of grain produced per kg N applied, increased in the order CC-int (62)oCS-int (74)oCC-rec (83)oCS-rec (123 kg kg 1), which is significantly higher than national averages of about 58–60 kg kg 1 in recent years (Dobermann & Cassman, 2002). Soil CO2 fluxes CO2 fluxes were low in March/April, but increased throughout May to July with increasing temperature and progressing growth of maize, reaching a maximum of 40–60 kg CO2-C ha 1 day 1 around silking stage (Fig. 1). Total soil CO2 efflux during the 2004 growing season was 20–40% lower than in the comparable period of 2003 (Fig. 1, Table 3), primarily due to cooler weather in 2004 and less need for irrigation (304 mm in 2004 vs. 445 mm in 2003). Soil CO2 fluxes in maize following maize were higher than in maize following soybean at most sampling dates (Fig. 1). In 2003 and 2004, average daily CO2 emission was 25.4 kg CO2-C ha 1 day 1 in CC-rec and CC-int as compared with 19.2 kg CO2-C ha 1 day 1 under maize in the CS-rec and CS-int systems (Po0.001). Total annual CO2 efflux from the soil surface was significantly larger for maize grown in CC systems than maize grown in the CS systems, with most of this difference occurring during the growing season (from emergence to harvest) (Table 3). Within each crop rotation and year, however, soil CO2 emissions were not significantly different between systems with recommended or intensive management during the growing season (Table 3). In the nongrowing season (fall to spring), differences caused by different management intensity were only significant in 2003, when systems with intensive r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 1978 M . A . A . A D V I E N T O - B O R B E et al. 50 Max. temperature Min. temperature Rain/ irrigation 70 20 50 10 40 0 P1M1 P3M2 30 −10 20 10 VE V6 70 VT PM VE V6 VT PM CC-rec CC-int 60 Air temperature (°C) 30 60 0 kg CO2-C ha−1 day−1 40 80 60 50 40 40 20 30 0 20 10 0 70 CS-rec CS-int 60 50 40 100 80 60 40 20 0 kg fertilizer N ha−1 Rainfall / irrigation (mm) 80 30 20 10 0 M 4 ay 04 Ju n 04 Ju l0 Au 4 g 0 Se 4 p 04 O ct 0 N 4 ov 04 Ap r 03 Ju l0 3 Au g 03 Se p 03 O ct 03 N ov 03 Ju n r0 M Ap ay 3 03 0 Fig. 1 Soil surface carbon dioxide (CO2) fluxes from continuous maize (CC, middle panel) and maize following soybean (CS, lower panel) at recommended (-rec) and intensive management levels (-int) during 2003–2004. Arrows and their location indicate date and amount of fertilizer N applications (kg N ha 1) in intensive treatments. VE, V6, VT and PM correspond to emergence, sixth leaf stage, tasseling and physiological maturity of maize, respectively. management tended to have somewhat higher CO2 fluxes than those with recommended management (Table 3). Likewise, whereas 64–67% of the total annual soil CO2 emission in CC-rec and CS-rec occurred during the growing season, this period accounted for only 55– 63% of the annual total in the intensively managed systems CC-int and CS-int (Table 3). Using the approach by Raich & Mora (2005), we estimated that during the growing season more than half of the CO2 produced was ascribed to autotrophic respiration (50–79%), (i.e. root respiration and fine-root turnover. On an annual basis, autotrophic respiration accounted for 24–44% of total soil respiration in all maize treatments and years measured. In 2005, soil CO2 fluxes were measured in maize (CC-rec and CC-int), the soybean phase of the maize– soybean rotation (CS-rec and CS-int), and in the adjacent grassland area. CO2 fluxes followed similar seasonal patterns, with seasonal peaks reached in mid-July at silking stage of maize or mid-pod elongation stage of soybean (Fig. 2). In the grassland site, soil CO2 emission was low in spring and autumn and highest in early June. Integrated over time, there was no significant difference in measured CO2 fluxes between maize and soybean during the 2005 growing season (Table 3). In the nongrowing period, however, soybean plots in CS-rec and CS-int had 34–42% lower CO2 losses than the CC systems (Table 3). Total annual soil CO2 emission in 2005 increased in the order soybean in CS-rec (4290 kg C ha 1)osoybean in CS-int (4980 kg C ha 1) omaize in CC-rec (5290 kg C ha 1)omaize in CC-int (5810 kg C ha 1)ograssland (6360 kg C ha 1). On an r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 S O I L G H G F L U X E S A N D G L O B A L WA R M I N G 1979 Table 3 Soils surface fluxes of CO2, N2O and CH4 in four maize-based cropping systems with recommended (-rec) and intensive (-int) management CO2-C (kg C ha 1)* Crop Growing seasonw Maize following maize Maize following soybean Soybean following maize Nongrowing seasonz Maize following maize Maize following soybean Soybean following maize Total annual flux Maize following maize Maize following soybean Soybean following maize N2O-N (kg N ha 1)* CH4-C (kg C ha 1)* System 2003 2004 2005 2003 2004 2005 CC-rec CC-int CS-rec CS-int CS-rec CS-int 5890a 5970a 3950b 4420b 3990a 3530ab 3150bc 2830c 3190ab 3270ab 2.99a 4.91a 3.38a 4.55a 2.31ab 2.48a 2.17b 1.82bc 1.30a 1.25a CC-rec CC-int CS-rec CS-int CS-rec CS-int 2930b 4200a 1920c 3560ab CC-rec CC-int CS-rec CS-int CS-rec CS-int 8820b 10170a 5870c 7980b 2910b 3510a 1940ab 2220a 1800b 1640b 2100b 2540a Grassland 5290bc 5810b 1.27a 1.61a 1.47a 1.01a 2004 1.68a 0.80a 2.19a 0.82a 0.30b 1.33a 0.93b 4.33a 1.13b 0.73b 0.10c 0.04c 0.27b 0.75a 1380c 1470c 5930a 5750ab 4950bc 4470c 2003 0.08a 0.55a 2.41a 2.52a 2.44a 2.57a 4290d 4980c 6363a 1.38a 1.80a 1.38a 1.90a 1.70a 0.96a 2.62b 4.16b 0.95a 0.61a 1.15a 1.69a 0.58a 0.30a 0.10a 0.16a 3.92c 9.24a 4.51bc 5.28b 2005 2.50a 1.86a 2.64a 1.51a 3.89b 5.76b 0.52a 0.33a 2.83a 2.49a 2.77a 1.11a 0.40b 1.49a 0.45b 3.88a 3.76a 4.34a 2.48a – *Within each column, season, and year emissions followed by the same letter are not significantly different at Po0.05. w Growing season refers to the period from emergence to harvest of the crop. Nongrowing season refers to the period from harvest of the crops in fall to crop emergence in spring. CO2, carbon dioxide; N2O, nitrous oxide. z annual basis, autotrophic respiration was estimated to account for 24–36% of total soil respiration in maize and soybean treatments, but was only 20% in the grassland. In a complete 2-year crop rotation with flux measurements conducted in maize and soybean (2004–2005), soil CO2 efflux in continuous maize systems was 22% larger than in maize–soybean rotations at both levels of management intensity (Table 4). Within each crop rotation, intensified management did not cause a significant increase in CO2 emissions as compared with the recommended best management practice. Soil N2O fluxes Daily N2O fluxes under maize grown in 2003 and 2004 ranged from –6 to 65 g N2O-N ha 1 day 1 (Fig. 3) and most of the total annual N2O loss occurred during the growing season (Table 3). Soil N2O fluxes were mostly low in late fall and early spring, but increased as the soil began to warm in late May. N2O fluxes in CC-int measured during the early spring period of 2003 were significantly higher than in the other three systems, but no such difference was observed in 2004 (Fig. 3). Residual soil NO3-N content in spring of 2003 was 172 kg N ha 1 in CC-int as compared with 67–75 kg N ha 1 in the other treatments (Table 2). The higher residual nitrate levels in CC-int resulted from applying fertilizer-N on the maize residue before plowing in the previous fall, which accelerated crop residue decomposition. In the spring of 2003, soil moisture levels and temperature (several days with 430 1C) were high, resulting in the increased N2O fluxes from the bare soil in CC-int. In contrast, in 2004, N2O fluxes in spring remained low in all treatments due to cool and drier spring weather (Fig. 3), even though soil nitrate levels measured in CC-int were similar to those in 2003 and also higher than in the other systems (Table 1). In 2003, N2O emissions during the reproductive period of maize (July to September) remained significantly higher in both CC-int and CS-int as compared with CC-rec and CS-rec (Fig. 3, P 5 0.0002). No such differences were measured in 2004 (P 5 0.989). N amounts (Table 2) and timing of applications (Fig. 3) were similar in both years. However, due to warmer r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 1980 M . A . A . A D V I E N T O - B O R B E et al. Soybean Rainfall/ irrigation (mm) Max. temperature Min. temperature Rain / irrigation 90 80 50 40 30 70 20 60 10 50 0 40 −10 30 20 Air temperature (°C) Maize 100 10 VE V6 VT PM CC-rec CC-int Grassland 120 100 VE R1 R3.5 PM CS-rec CS-int 100 80 80 60 60 40 40 20 20 kg fertilizer N ha−1 g N2O-N ha−1 day−1 0 140 0 0 kg CO2-C ha−1 day−1 60 50 40 30 20 10 Ap r M 05 ay 0 Ju 5 n 0 Ju 5 l0 Au 5 g 0 Se 5 p 0 O 5 ct 0 N 5 ov D 05 ec 05 Ap r M 05 ay 0 Ju 5 n 0 Ju 5 l0 Au 5 g 0 Se 5 p 0 O 5 ct 0 N 5 ov D 05 ec 05 0 Fig. 2 Soil carbon dioxide (CO2) and nitrous oxide (N2O) fluxes from continuous maize (CC, left) and soybean following maize (CS, right) at recommended (-rec) and intensive management levels (-int) during 2005. Arrows and their location indicate date and amount of fertilizer N application (kg N ha 1) in intensive treatments. VE, V6, VT, R1, R3.5 and PM correspond to emergence, sixth leaf stage, tasseling (maize), first flowering (soybean), mid-pod (soybean) and physiological maturity, respectively. weather in the 2003 growing season, total growing season evapotranspiration in 2003 was 741 mm as compared with 649 mm in 2004, resulting in more frequent sprinkler irrigations from early July to midSeptember. The elevated late-season N2O fluxes in 2003 (30–60 g N2O-N ha day 1) in CC-int and CS-int occurred despite low soil nitrate levels in the top 15 cm of soil (Fig. 4). They coincided primarily with frequent irrigation and occasional rainfall events (Fig. 4) and 455% WFPS in the surface soil layer. During the same period in 2004, cooler weather and lessfrequent rainfall and irrigation events prevailed and N2O fluxes were consistently below 30 g N2ON ha 1 day 1 (Fig. 3). With the exception of CC-int in 2003, cumulative soil N2O emissions under maize were generally higher during the growing seasons than in the nongrowing seasons in all years. Total N2O emissions under maize were higher in intensive treatments than in recommended ones in 2003, but the difference was only statistically significant for CC-int vs. CC-rec (Table 3). In 2004, N2O emissions remained low during both growing and nongrowing seasons and the differences among the four cropping systems were not significant r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 1981 S O I L G H G F L U X E S A N D G L O B A L WA R M I N G Table 4 Average annual emissions (or uptake) of CO2, N2O and CH4 measured at the soil surface for one complete crop rotation cycle (2004–2005) System Continuous maize CC-rec CC-int Maize–soybean CS-rec CS-int CO2 (kg CO2-C ha 1 yr 1) N2O (kg N2O-N ha 1 yr 1) CH4 (kg CH4-C ha 5612a 5779a 1.87a 2.16a 3.36a 3.13a 4621b 4726b 1.42b 2.03a 3.56a 1.80a 1 yr 1) Within each column, means followed by the same letter are not significantly different at Po0.05. CO2, carbon dioxide; N2O, nitrous oxide. (Table 3). Measurements conducted shortly before and after N applications did not indicate large increases in N2O fluxes following N application events (Fig. 3). In 2005, soil N2O fluxes were measured in maize (CCrec and CC-int), in the soybean phase of the maize– soybean rotation (CS-rec and CS-int), and in the adjacent grassland area. A large and highly variable N2O emission peak was observed under maize in both CC-rec and CC-int on June 6, 2005 (50–140 g N2ON ha 1 day 1). On that day, maximum air temperature was 32 1C and both WFPS (460%) and soil NO3-N (59– 61 mg NO3-N kg 1) measured in the top 0.15 m were high in these plots. The increased soil nitrate levels resulted from preplant N application in late April (160 kg N ha 1 in CC-rec, 100 kg N ha 1 in CC-int) and very little crop N uptake up to that date. However, despite this short-lived peak, N2O fluxes under maize remained below those measured in 2003 and 2004 and there was no difference between the CC-rec and CC-int treatments (Table 3, Fig. 2). Low levels of soil NO3-N in 2005 probably contributed to lower N2O fluxes measured. In both continuous maize systems, initial soil NO3-N content in spring was 45–79% less than in previous years because 2004 had been a high-yielding year with large crop N removal (Table 2). In the CS-rec system with no N application to soybean, N2O emissions remained low throughout the growing season. Late application of N fertilizer to soybean in CS-int followed by irrigation did not significantly increase crop yield (Table 2), but caused an increase in N2O emissions during the reproductive period (Fig. 2). N2O emissions in the grassland were low and similar to the N2O emissions in the unfertilized soybean crop in CS-rec (Fig. 2). Total annual N2O emission in 2005 was low (Table 3), ranging from 0.40 kg N2O-N ha 1 for grassland and soybean in CS-rec to 1.38–1.80 kg N2O-N ha 1 in other treatments. Averaged across a 2-year crop rotation cycle, cumulative soil N2O emissions were lowest in CS-rec, whereas no significant differences occurred between the other three cropping systems (Table 4). In general, however, annual average N2O emissions were low in all four systems (1.4–2.2 kg N2O-N ha 1 yr 1) because the crop and fertilizer management practices employed resulted in high crop N uptake (293–380 kg N ha 1 yr 1) and, thus, high fertilizer-N recovery efficiency. Environmental variables related to soil CO2 and N2O emissions Nitrous oxide fluxes were positively correlated with soil respiration, suggesting that both were controlled to some extent by similar environmental factors. This relationship was stronger in the nongrowing season (r 5 0.48; Po0.001) than in the growing season (r 5 0.40; Po0.001), probably due to increasing relative contributions of autotrophic respiration to the total soil CO2 efflux measured. Generally, only air temperature, soil temperature, WFPS, and the 5- or 3-day cumulative rainfall and irrigation amount before a flux measurement had significant correlations with CO2 and N2O fluxes in all cropping systems. Soil temperature was positively correlated with CO2 and N2O fluxes in all three depths (CO2: r 5 0.39–0.63, Po0.001; N2O: r 5 0.25–0.58, Po0.001) while WFPS was correlated with GHG fluxes only at 0–0.05 m soil depth in 2003 and 2004 (r 5 0.27–0.45; Po0.001). Several high N2O fluxes were measured at 60–90% WFPS. The 3- or 5-day cumulative rainfall and irrigation amounts before a flux measurement correlated positively with both CO2 and N2O fluxes in 2003 and 2004 (r 5 0.43–0.54, Po0.001). Correlations were generally poorer in 2005 due to the lower level of gas fluxes in that year. Unlike earlier findings by Amos et al. (2005), there was no significant correlation between pH or soil EC and the fluxes of CO2 or N2O (data not shown). Lime applications and changing from drip irrigation (1999– 2002) to sprinkler irrigation (2003–2005) ensured optimal pH ranges (6 to 6.5) and more uniform nutrient and water distribution and thus also lower EC (0.2–0.9 dS m 1) in the topsoil. Likewise, correlations between soil NO3-N concentration in three depths and fluxes of r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 1982 M . A . A . A D V I E N T O - B O R B E et al. 70 50 Max. temperature Min. temperature Rain/irrigation 40 30 60 20 50 10 40 0 30 −10 20 Air temperature (°C) Rainfall/ irrigation (mm) 80 10 0 VT PM VE V6 PM CC-rec CC CC-int 60 g N2O-N ha−1 day−1 VT 100 80 60 40 40 20 20 0 0 100 CS-rec CS CS-int 80 80 60 60 40 40 20 20 0 l0 Au 4 g 0 Se 4 p 0 O 4 ct 0 No 4 v 04 04 n Ju Ju 4 04 r0 ay M Ap 03 l0 Au 3 g 0 Se 3 p 03 O ct 0 No 3 v 03 Ju 03 n Ju ay M r0 3 0 Ap kg fertilizer N ha−1 VE V6 80 Fig. 3 Soil surface nitrous oxide (N2O) fluxes from continuous maize (CC, middle panel) and maize following soybean (CS, lower panel) at recommended (-rec) and intensive management levels (-int) during 2003–2004. Arrows and their location indicate date and amount of fertilizer N applications (kg N ha 1) in intensive treatments. VE, V6, VT and PM correspond to emergence, sixth leaf stage, tasseling and physiological maturity of maize, respectively. CO2 or N2O were generally not significant (Fig. 5). Occasional peaks in N2O fluxes were not always related to high soil NO3-N levels (Figs 4 and 5). The correlation between N2O flux and soil nitrate was particularly poor under warm (415 1C soil temperature) and wet (460% WFPS) conditions (r 5 0.05, ns) when most of the peak N2O fluxes occurred, often at relatively low soil NO3-N (Fig. 5). Small negative fluxes ( 6 to 0.29 g N2O-N ha 1 day 1) were observed, particularly in the recommended cropping systems in early spring and late autumn (Figs 1–4). Some negative fluxes occurred under conditions that have been proposed to favor the use of atmospheric N2O as the only electron acceptor left for denitrification, (i.e. wet soil, very low soil NO3-N concentration and moderate temperatures (Ryden, 1981). However, small negative fluxes also occurred in drier soil or at higher NO3 levels (Fig. 4), suggesting the presence of other N2O uptake mechanisms (Flechard et al., 2005). Soil organic C and GWP Since the start of this experiment in 1999, large amounts of crop residue have been returned to the soil in all four management systems, but with significant differences r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 S O I L G H G F L U X E S A N D G L O B A L WA R M I N G 120 80 60 50 CC-rec CC-int CS-rec CS-int 40 30 20 10 Soil NO3-N content (g N kg−1) Rainfall and irrigation (mm) 0 70 Tsoil <15°C WFPS <60% Tsoil >15°C WFPS <60% Tsoil <15°C WFPS >60% 100 g N2O-N ha−1 day−1 g N2O-N ha−1 day−1 70 1983 Rainfall and irrigation 60 80 60 r = 0.31 r = 0.48 r = 0.05 40 20 0 −20 50 Tsoil >15°C WFPS >60% r = 0.32 0 10 20 30 40 50 60 70 80 Soil NO3-N content (mg N kg−1) 40 Fig. 5 Relationship between nitrous oxide flux and soil nitrate content in 0–0.15 m soil depth for different categories of soil temperature (Tsoil) and water-filled pore space (WFPS). Values shown are measurements conducted in 2003 and 2005. 30 20 10 0 18 16 14 12 10 8 6 4 2 0 60 80 100 120 140 160 180 200 220 240 260 280 300 Day of year Fig. 4 Dynamics of soil nitrate content in 0–0.15 m depth, rainfall/irrigation events, and nitrous oxide (N2O) fluxes during March to October 2003. among them (Fig. 6). Relative to the CS-rec system, average annual C return with aboveground residue increased in the order CS-recoCS-int ( 1 8%)oCC-rec ( 1 22%)oCC-int ( 1 39%). Consequently, both SOC and TSN increased in the two CC systems, but decreased in CS-rec or remained unchanged in CS-int (Fig. 6). On average, SOC declined by 30 g C m 2 yr 1 in CS-rec, whereas it increased at a rate of 62 g C m 2 yr 1 in CCint (Table 5). The different changes in SOC largely reflected the increasing crop residue input (Fig. 6) and, within each crop rotation, the lack of significant differences in soil CO2 fluxes between recommended and intensively managed systems (Tables 3 and 4). It should be noted, however, that due to spatial variability the standard errors of the cumulative changes in SOC measured for the 2000–2005 period were relatively large ( 60–130 g C m 2, Fig. 6). More experimental years will be needed to verify the initial trends observed. With conventional use of maize and soybean grain as feed and food, all four cropping systems where net sources of GHG, with GWP ranging from 54 to 102 g CO2-C m 2 yr 1 (Table 5). Positive or negative changes in SOC, intrinsic C costs associated with crop production and soil N2O emissions were major contributors to the net GWP, whereas CH4 oxidation added only little mitigation capacity. N fertilizer (16–36%), energy used for irrigation (15–22%), electricity for grain drying (13–18%), diesel (10–16%) and lime (9–13%) were the major components of the C costs associated with the agricultural production. Despite higher C cost associated with crop production and also higher N2O emissions, net GWP in continuous maize systems was lower than that of the maize–soybean systems because sequestration of atmospheric CO2 in SOC was only observed in CC-rec ( 44 g C m 2 yr 1) and CC-int ( 62 g C m 2 yr 1) (Table 5). Within each crop rotation, intensification of management practices increased production C costs and also N2O emissions, but when combined with the net change in SOC resulted in only slightly higher GWP for CC-int as compared with CCrec or no change in GWP for CS-int as compared with CS-rec (Table 5). Large variations in N2O emissions among years caused, however, large interannual variation in the GWP of these systems. In the CC-rec system, r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 1984 M . A . A . A D V I E N T O - B O R B E et al. Residue C Residue N a a a b 11 c 500 d 10 b 9 b 400 8 300 7 CS-rec Cumulative change in SOC (g Cm−2) 12 600 CS-int CC-rec 500 400 (b) a 50 40 a a 30 20 200 b 100 10 0 0 −100 b −10 b −20 b −300 CS-rec CS-int CC-rec CC-int −30 less than those of fossil-fuel based gasoline. In our study, using the same calculation approach but with higher crop yields and also accounting for soil C sequestration, the net reduction in life cycle GHG emissions of maize ethanol relative to petrol-based gasoline was 38% in CC-rec and 33% in CC-int. Similar to the findings of Farrell et al. (2006) agricultural production accounted for 33–37% (continuous maize) or 42–44% (for maize in the maize–soybean systems) of the life cycle GHG emissions of maize ethanol. Discussion CC-int a ∆SOC ∆TSN 300 −200 Annual residue N (g m−2 yr −1) 13 (a) Cumulative change in TSN (g N Cm−2) Annual residue C (g m−2 yr−1) 700 Fig. 6 Average annual input of carbon (C) and nitrogen (N) with aboveground crop residues during 1999–2004 (a) and changes in soil organic C (SOC) and total soil N (TSN) for the corresponding period (b) in continuous maize (CC) and maize– soybean (CS) systems at Lincoln, Nebraska. Means and standard errors of two population densities for each management system. Letters indicate statistical significance (Po0.05) of treatment differences (Holm–Sidak test). for example, annual N2O emissions ranged from 1.28 to 3.92 kg N2O-N ha 1 (Table 4), which is equivalent to a GWP range of 32 g CO2-C m 2 yr 1. In the CC-int system, annual N2O emission ranged from a low of 1.8 kg N2O-N ha 1 in 2005 to high of 9.24 kg N2O-N ha 1 in 2003 (Table 4), which is equivalent to a GWP range of 94 g CO2-C m 2 yr 1. Low N2O fluxes in 2004 and 2005 illustrate, however, that this range can be narrowed down significantly with good crop management. Owing to potential offset of fossil fuel CO2, use of grain for biofuel production represents an additional mitigation option for lowering the GHG emissions from agricultural systems. Farrell et al. (2006) concluded that at current average maize production and ethanol conversion efficiency levels, GHG emissions from corn ethanol (including agricultural production, biorefinery, coproduct credits and transport) are only about 13% Our results agree with previous studies showing that the primary controls on soil CO2 fluxes include temperature, water content, crop residue amount and quality, and plant C allocation to roots (Rochette & Flanagan, 1997; Smith et al., 2003; Reth et al., 2005). Despite the large biomass production in our high-yielding maize systems, peak growing season (about 40– 60 kg C ha 1 day 1, Figs 1 and 2) or annual (4.3– 10.2 Mg C ha 1 yr 1, Table 3) soil CO2 efflux was within typical ranges for arable crops (Alvarez et al., 1995; Kessavalou et al., 1998; Hu et al., 2001). Significant potential for sequestration of atmospheric C exists in intensively managed continuous maize systems. In CC-int, 14% more crop residue C was returned to the soil than in CC-rec, but there was no significant difference in soil CO2 fluxes (Fig. 1, Table 4). Likewise, residue C amounts in CC-int were 28–39% larger than in the two maize–soybean systems as compared to soil CO2 efflux differences of only 22% to 25%. Measured changes in SOC (Fig. 6) confirmed the high C sequestration potential in the two continuous maize systems, thus highlighting the importance of biomass production and crop residue management for soil C sequestration. In contrast, loss of SOC and TSN occurred in the more commonly practiced maize–soybean system (CS-rec) because of high crop N removal with grain (Table 1), less residue input and fast turnover of relatively N-rich soybean residue after incorporation in the soil. Our results (Fig. 6) confirm those of eddy covariance studies at other sites, showing that significant net C losses during the soybean phase limit the soil C sequestration potential in maize–soybean rotations of North America (Baker & Griffis, 2005; Verma et al., 2005). Applying N fertilizer in fall on maize residue (CC-int since 2001; CS-int since 2004) followed by relatively deep but noninverting incorporation probably enhanced the formation of more stable humus compounds resulting from residue decomposition during the fall to spring period. This seems to contradict the widespread notion that conservation tillage is required for r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 S O I L G H G F L U X E S A N D G L O B A L WA R M I N G 1985 Table 5 Estimated net global warming potential (GWP) in maize-based cropping systems with recommended and intensive management GWP components Agricultural production* N fertilizer P, K, fertilizer Lime Seed, pesticides Machinery, transport Diesel Irrigation Grain drying Total D Soil Cw Soil N2Oz Soil CH4z GWP§ Continuous maize (CC) (g CO2-C equivalents m 2 yr 1) Maize–soybean (CS) (g CO2-C equivalents m 2 yr 1) Recommended Recommended Intensive 8 0 6 5 2 8 11 9 49 30 25 2 102 18 6 9 6 3 8 11 10 71 2 34 1 102 22 0 6 5 2 9 14 11 69 44 32 3 54 Intensive 33 6 9 6 3 9 14 12 92 62 57 3 84 *Carbon cost associated with crop production. Average for maize and soybean crops grown during 2000–2005. w Average annual change in SOC, based on measurements of SOC conducted in June 2000 and June 2005. Radiative forcing potential for 100-year time frame. CC systems: average of three maize crops (2003–2005); CS systems, weighted average of two maize crops (2003–2004) and one soybean crop (2005). § GWP 5Agricultural production 1 DSOC 1 soil N2O 1 soil CH4. z sequestering atmospheric CO2 in agricultural soils (West & Post, 2002; Lal et al., 2003). Recent studies suggest, however, that when sampling is done deep enough and SOC stocks are properly expressed on an equivalent soil dry mass basis, the potential for no-till systems to sequester atmospheric CO2 in SOC seems limited (VandenBygaart & Angers, 2006; Baker et al., 2007). Particularly in high-yielding systems with large amounts of crop residue, no-till cropping is not necessarily the best management strategy. Verma et al. (2005) used both soil C and eddy-covariance measurements to study CO2 fluxes and C budgets in irrigated and rainfed maize-based agroecosystems under no-till in eastern Nebraska. The lack of net C sequestration in their study was primarily due to high ecosystem respiration losses, including the decomposing residue on the soil surface. With the exception of the CC-int system in 2003, annual N2O emissions under maize were within 1.4– 5.3 kg N2O-N ha 1 yr 1 (Table 3), which is at the lower end of values reported for similar studies in North America (Thornton & Valente, 1996; Qian et al., 1997; Ginting & Eghball, 2005; Venterea et al., 2005). In all systems studied, the lack of continuous N2O flux measurements, particularly during freeze-thaw periods, may have underestimated the annual N2O losses in some years. However, more important is the finding that although the amount of fertilizer N applied to maize grown in the intensive cropping systems was 40% (CC) or 64–92% (CS) greater than in the recommended cropping systems, N2O losses were not directly related to the level of N input only. This seems to contradict the assumptions made in the current IPCC method, which calculates the contributions of N fertilizer to global anthropogenic N2O fluxes by assuming that on average 1.25  1% of the N amount applied is lost as N2O (IPCC, 2001). More recent summaries of available measurements suggest that fertilizer-induced N2O emissions (FIE) from agricultural fields range from o0.5% to more than 2% of the fertilizer-N applied, with an estimated global average of 0.9% (Bouwman et al., 2002a, b; Stehfest & Bouwman, 2006). Background emissions of N2O with zero N application were only measured under soybean in the CS-rec system and in the grassland area in 2005, both amounting to 0.4 kg N2ON ha 1 yr 1. Assuming similar background emissions in other years, the FIE factor of N applied to maize ranged from 1.9% to 3.5% in 2003, from 0.8% to 1.5% in 2004 and from 0.4% to 0.5% in 2005, with no consistent differences among the four systems. The apparently higher FIE factors in 2003 are probably incorrect because actual background N2O emissions were likely to be higher in 2003 than in 2004 and 2005 due to significant soil NO3 carry-over and more frequent irrigation (Table 2). Soil N2O fluxes depend on environment  management interactions that influence the balance and rates of r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 1986 M . A . A . A D V I E N T O - B O R B E et al. microbial nitrification and denitrification processes and the transport of N2O (Smith et al., 2003). Temperature, moisture, pH, osmotic stress caused by soluble salts, supply of C and N compounds, and competition for mineral N by other sinks such as crops are key drivers of soil N2O fluxes (Weier et al., 1993; Skiba et al., 1998; Adviento-Borbe et al., 2006; Lee et al., 2006). Many of these factors can be manipulated through management practices such as tillage (Venterea et al., 2005), irrigation (Qian et al., 1997), N fertilizer source, timing and placement (Smith et al., 1997), or site-specific prescription of N fertilizer that accounts for differences in crop N demand (Sehy et al., 2003). Of particular interest in our study were the relatively small treatment differences in N2O fluxes despite the large differences in management systems and N inputs (Table 4) and the poor relationship between N2O emissions and soil NO3-N levels (Fig. 5). McSwiney & Robertson (2005) have shown how, with other factors held constant, increasing the fertilizer dose increases the concentrations of NO3-N in surface soil and also N2O fluxes. In their study, the corresponding increase in N2O was not linearly related to the N rate, (i.e. it became nonlinear when the N rate exceeded the amount required to meet crop N demand. Figures 4 and 5 were drawn to be directly comparable with Figs 5 and 6 in McSwiney & Robertson (2005). At similar levels of fertilizer-N input, both soil NO3-N and N2O fluxes in our study were nearly a magnitude lower than those reported in their study and the correlation between soil NO3-N and measured N2O fluxes depended mostly on soil temperature and water content. The major difference between these two studies lies in the general level of crop management and the crop yields, which in our experiment were about twice as high than those reported by McSwiney & Robertson (2005). Major management differences included optimization of planting dates and hybrid choice to maximize crop yield potential, higher plant populations, site-/season-specific prescription of N rates, N application methods that aimed at avoiding excessive NO3-N levels near the soil surface, and irrigation. Hence, N2O emissions are likely to be more related to soil N turnover rather than mineral N pool size per se (Mosier et al., 1996). Managing the crop N sink is an important mitigation option because it directly affects the economic and ecological performance of an agricultural system. Our results reiterate the need for complete GWP accounting in studies on GHG emissions and soil sequestration. The GWP estimates of the four systems studied (Table 5) are associated with large uncertainties about the changes in SOC (Fig. 6), the GHG fluxes and to what extent large, climate-dependent interannual fluctuations in N2O emissions could be controlled through better management. Hence, Table 5 must be treated with caution and is meant to illustrate the major differences among the four cropping systems, indicating that optimized management at high-yield levels can have positive effects on the GWP of intensively managed agricultural systems. This needs to be confirmed through detailed measurement under production scale conditions. Major components for improving crop management to reduce GWP are (i) choosing the right combination of adopted varieties, planting date and plant population to maximize yield potential, crop biomass productivity and residue input, (ii) tactical water and N management decisions that minimize energy use, achieve high N use efficiency and avoid high N2O emissions and (iii) a tillage and residue management approach that can handle the large amounts of residue produced and favors the build-up of soil organic matter. Managing a crop at high yield levels creates large sinks for CO2 and mineral N, thereby providing the prerequisite for sequestering atmospheric CO2 and avoiding large N2O emissions that could results from inefficient utilization of soil or fertilizer N. Future research should concentrate on demonstrating the potential impact of such management practices at production scales, particularly whether it is possible to reduce the large seasonal fluctuations in N2O and CO2 emissions from the soil surface. Reduced N2O losses may not be the only or most important environmental consequence. Direct measurements of nitrate leaching and its impact on water quality were beyond the scope of our study, but it is likely that the management practices employed resulted in both reduced N2O and NO3 leaching losses. The relative environmental benefits associated with this must be quantified in future studies. Conclusions Seasonal variations in soil CO2 and N2O fluxes were principally dependent upon temperature, soil water status associated with precipitation and irrigation events, crop growth, and, to a lesser extent soil NO3-N content. CO2 emissions were greatly influenced by temperature and the amount and kind of crop residue returned to the field from the previous crop. Peaks in N2O primarily occurred when high soil/air temperatures coincided with wet soil conditions, irrespective of soil nitrate levels. Intensification of cropping does not necessarily increase GHG emissions and GWP of agricultural systems provided that crops are grown with best management practices and near yield potential levels, resulting in high resource use efficiency. Continuous maize systems r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 S O I L G H G F L U X E S A N D G L O B A L WA R M I N G had lower net GWP than maize–soybean systems, primarily due to greater crop residue amounts and the use of a deep tillage and residue management practice that favored the build-up of soil organic matter. The potential for GHG mitigation is further increased when corn is converted to bioethanol. Policies that favor greater adoption of resource-efficient management practices for closing existing yield gaps would not only satisfy the increasing demands for crops such a maize and soybean, but may also mitigate GHG emissions from agriculture. Acknowledgements We thank Greg Teichmeier, Timothy McAndrew, Brigid Amos, Kenneth Cassman, Tim Arkebauer, James Specht, Haishun Yang, Adam Liska, John Doran, Steve Comfort and Rhae Drijber (all University of Nebraska, Lincoln) for their contributions to this work. This research was supported through funds provided by the International Plant Nutrition Institute (IPNI), the Foundation for Agronomic Research (FAR), the Fluid Fertilizer Foundation (FFF), the US Department of Energy, EPSCoR program, Grant No. DE-FG-02-00ER45827, and the Hatch Act. References Adviento-Borbe MAA, Doran JW, Drijber RA, Dobermann A (2006) Soil electrical conductivity and water content affect nitrous oxide and carbon dioxide emissions in intensively managed soils. Journal of Environmental Quality, 35, 1999–2010. Alvarez R, Santanatoglia OJ, Garcia R (1995) Soil respiration and carbon inputs from crops in a wheat soybean rotation under different tillage systems. Soil Use and Management, 11, 45–50. Amos B, Arkebauer TJ, Doran JW (2005) Soil surface fluxes of greenhouse gases in an irrigated maize-based agroecosystem. Soil Science Society of America Journal, 69, 387–395. Baker JM, Griffis TJ (2005) Examining strategies to improve the carbon balance of corn/soybean agriculture using eddy covariance and mass balance techniques. Agricultural and Forest Meteorology, 128, 163–177. Baker JM, Ochsner TE, Venterea RT, Griffis TJ (2007) Tillage and soil carbon sequestration – What do we really know? Agriculture, Ecosystems and Environment, 118, 1–5. Bouwman AF, Boumans LJM, Batjes NH (2002a) Emissions of N2O and NO from fertilized fields: summary of available measurement data. Global Biogeochemical Cycles, 16, 1058, doi: 10.1029/2001GB001811. Bouwman AF, Boumans LJM, Batjes NH (2002b) Modeling global annual N2O and NO emissions from fertilized fields. Global Biogeochemical Cycles, 16, 1080, doi: 10.1029/ 2001GB001812. Cassman KG, Dobermann A, Walters DT, Yang HS (2003) Meeting cereal demand while protecting natural resources and improving environmental quality. Annual Review of Environment and Resources, 28, 315–358. Dobermann A, Cassman KG (2002) Plant nutrient management for enhanced productivity in intensive grain production systems of the United States and Asia. Plant and Soil, 247, 153–175. 1987 Fahey T, Tierney GL, Fitzhugh RD, Wilson GF, Siccama TG (2005) Soil respiration and soil carbon balance in a northern hardwood forest ecosystem. Canadian Journal of Forest Research, 35, 244–253. Farrell AE, Plevin RJ, Turner BT, Jones AD, O’Hare M, Kammen DM (2006) Ethanol can contribute to energy and environmental goals. Science, 311, 506–508. Flechard CR, Neftel A, Jocher M, Amman C, Fuhrer J (2005) Bidirectional soil/atmosphere N2O exchange over two mown grassland systems with contrasting management practices. Global Change Biology, 11, 2114–2127. Flessa H, Ruser R, Schilling R, Loftfield N, Munch JC, Kaiser EA, Beese F (2002) N2O and CH4 fluxes in potato fields: automated measurement, management effects and temporal variation. Geoderma, 105, 307–325. Gifford RM, Roderick ML (2003) Soil carbon stocks and bulk density: spatial or cumulative mass coordinates as a basis of expression? Global Change Biology, 9, 1507–1514. Ginting D, Eghball B (2005) Nitrous oxide emission from no-till irrigated corn: temporal fluctuation and wheel traffic effects. Soil Science Society of America Journal, 69, 915–924. Hu R, Kusa K, Hatano R (2001) Soil respiration and methane flux in adjacent forest grassland and cornfield soils in Hokkaido, Japan. Soil Science and Plant Nutrition, 47, 621–627. IPCC (2001) Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK. Kessavalou A, Mosier AR, Doran JW, Drijber RA, Lyon DJ, Heinemeyer O (1998) Fluxes of carbon dioxide, nitrous oxide, and methane in grass sod and winter wheat-fallow tillage management. Journal of Environmental Quality, 27, 1094–1104. Lal R, Follett RF, Kimble JM (2003) Achieving soil carbon sequestration in the United States: a challenge to the policy makers. Soil Science, 168, 827–845. Lee J, Six J, King AP, van Kessel C, Rolston DE (2006) Tillage and field scale controls on greenhouse gas emissions. Journal of Environmental Quality, 35, 714–725. McSwiney CP, Robertson GP (2005) Nonlinear response of N2O flux to incremental fertilizer addition in a continuous maize (Zea mays L.) Cropping system. Global Change Biology, 11, 1712–1719. Mosier AR, Duxbury JM, Freney JR, Heinemeyer O, Minami K (1996) Nitrous oxide emissions from agricultural fields: assessment, measurement and mitigation. Plant and Soil, 181, 95–108. Norman JM, Garcia R, Verma SB (1992) Soil surface CO2 fluxes and the carbon budget of a grassland. Journal of Geophysical Research, 97, 845–853. Qian JH, Doran JW, Weier KL, Mosier AR, Peterson TA, Power JF (1997) Soil denitrification and nitrous oxide losses under corn irrigated with high-nitrate groundwater. Journal of Environmental Quality, 26, 348–360. Raich JW, Mora G (2005) Estimating root plus rhizosphere contributions to soil respiration in annual croplands. Soil Science Society of America Journal, 69, 634–639. Reth S, Reichstein M, Falge E (2005) The effect of soil water content, soil temperature, soil pH-value and the root mass on soil CO2 efflux. A modified model. Plant and Soil, 268, 21–33. r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988 1988 M . A . A . A D V I E N T O - B O R B E et al. Robertson GP, Grace PR (2004) Greenhouse gas fluxes in tropical and temperate agriculture: the need for a full-cost accounting of global warming potentials. Environment, Development and Sustainability, 6, 51–63. Robertson GP, Paul EA, Harwood RR (2000) Greenhouse gases in intensive agriculture: contributions of individual gases to the radiative forcing of the atmosphere. Science, 289, 1922–1925. Rochette P, Flanagan LB (1997) Quantifying rhizosphere respiration in a corn crop under field conditions. Soil Science Society of America Journal, 61, 466–477. Ryden JC (1981) Nitrous oxide exchange between a grassland soil and the atmosphere. Nature, 292, 235–237. Sehy U, Ruser R, Munch JC (2003) Nitrous oxide fluxes from maize fields: relationship to yield, site-specific fertilization, and soil conditions. Agriculture, Ecosystems and Environment, 99, 97–111. Setiyono TD, Weiss A, Specht J, Bastidas AM, Cassman KG, Dobermann A (2007) Understanding and modeling the effect of temperature and daylength on soybean phenology under high-yield conditions. Field Crops Research, 100, 257–271. Shapiro CA, Ferguson RB, Hergert GW, Dobermann A, Wortmann CS (2003) Fertilizer Suggestions for Corn. NebGuide G74-174-A. Cooperate Extension, Institute of Agriculture and Natural Resources, University of Nebraska-Lincoln, Lincoln, NE. Skiba UM, Sheppard LJ, MacDonald J, Fowler D (1998) Some key environmental variables controlling nitrous oxide emissions from agricultural and semi-natural soils in Scotland. Atmospheric Environment, 32, 3311–3320. Smith KA, Ball T, Conen F, Dobbie KE, Massheder J, Rey R (2003) Exchange of greenhouse gases between soil and atmosphere: interaction between soil physical factors and biological process. European Journal of Soil Science, 54, 791. Smith KA, Mctaggart IP, Tsuruta H (1997) Emissions of N2O and NO associated with nitrogen fertilization in intensive agriculture, and the potential for mitigation. Soil Use and Management, 13, 296–304. Specht JE, Hume DJ, Kumudini SV (1999) Soybean yield potential – a genetic and physiological perspective. Crop Science, 39, 1560–1570. Stehfest E, Bouwman L (2006) N2O and NO emission from agricultural fields and soils under natural vegetation: summarizing available measurement data and modeling of global annual emissions. Nutrient Cycling in Agroecosystems, 74, 207–228. Thornton FC, Valente RJ (1996) Soil emissions of nitric oxide and nitrous oxide from no-till corn. Soil Science Society of America Journal, 60, 1127–1133. VandenBygaart AJ, Angers DA (2006) Towards accurate measurements of soil organic carbon stock change in agroecosystems. Canadian Journal of Soil Science, 86, 465–471. Venterea RT, Burger M, Spokas KA (2005) Nitrous oxide and methane emissions under varying tillage and fertilizer management. Journal of Environmental Quality, 34, 1467– 1477. Verma SB, Dobermann A, Cassman KG et al. (2005) Annual carbon dioxide exchange in irrigated and rainfed maize-based agroecosystems. Agricultural and Forest Meteorology, 131, 77–96. Weier KL, Doran JW, Power JF, Walters DT (1993) Denitrification and the dinitrogen/nitrous oxide ratio as affected by soil water, available carbon, and nitrate. Soil Science Society of America Journal, 57, 66–72. West TO, Post WM (2002) Soil organic carbon sequestration rates by tillage and crop rotation: a global data analysis. Soil Science Society of America Journal, 66, 1930–1946. Yang HS, Dobermann A, Lindquist JL, Walters DT, Arkebauer TJ, Cassman KG (2004) Hybrid-Maize – a maize simulation model that combines two crop modeling approaches. Field Crops Research, 87, 131–154. r 2007 The Authors Journal compilation r 2007 Blackwell Publishing Ltd, Global Change Biology, 13, 1972–1988