TECHNICAL REPORTS: ATMOSPHERIC POLLUTANTS AND TRACE GASES
Nitrogen, Tillage, and Crop Rotation Effects on Carbon Dioxide and Methane Fluxes from Irrigated Cropping Systems Francesco Alluvione University of Turin Ardell D. Halvorson* and Stephen J. Del Grosso USDA-ARS Long-term effects of tillage intensity, N fertilization, and crop rotation on carbon dioxide (CO2) and methane (CH4) flux from semiarid irrigated soils are poorly understood. We evaluated effects of: (i) tillage intensity [no-till (NT) and conventional moldboard plow tillage (CT)] in a continuous corn rotation; (ii) N fertilization levels [0–246 kg N ha–1 for corn (Zea mays L.); 0 and 56 kg N ha–1 for dry bean (Phaseolus vulgaris L.); 0 and 112 kg N ha–1 for barley (Hordeum distichon L.)]; and (iii) crop rotation under NT soil management [corn-barley (NTCB); continuous corn (NT-CC); corn-dry bean (NT-CDb)] on CO2 and CH4 flux from a clay loam soil. Carbon dioxide and CH4 fluxes were monitored one to three times per week using vented nonsteady state closed chambers. No-till reduced (14%) growing season (154 d) cumulative CO2 emissions relative to CT (NT: 2.08 Mg CO2–C ha–1; CT: 2.41 Mg CO2–C ha–1), while N fertilization had no effect. Significantly lower (18%) growing season CO2 fluxes were found in NT-CDb than NT-CC and NT-CB (11.4, 13.2 and 13.9 kg CO2–C ha–1d–1 respectively). Growing season CH4 emissions were higher in NT (20.2 g CH4 ha–1) than in CT (1.2 g CH4 ha–1). Nitrogen fertilization and cropping rotation did not affect CH4 flux. Implementation of NT for 7 yr with no N fertilization was not adequate for restoring the CH4 oxidation capacity of this clay loam soil relative to CT plowed and fertilized soil.
Copyright © 2009 by the American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America. All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Published in J. Environ. Qual. 38:2023–2033 (2009). doi:10.2134/jeq2008.0517 Received 15 Dec. 2008. *Corresponding author (
[email protected]). © ASA, CSSA, SSSA 677 S. Segoe Rd., Madison, WI 53711 USA
C
onversion of forest and grass ecosystems to rainfed and irrigated cropping systems has altered C cycle fluxes between soil and atmosphere by reducing the amount of residue returned to soil and increasing oxidation of soil organic matter (SOM) (Lal, 2004). Consequences of this conversion have been a depletion of soil fertility (Johnston, 1991; Bauer and Black, 1994; Doran and Parkin, 1994) and an increase in atmospheric CO2 (Houghton et al., 1983; Houghton, 1995). Adoption of agricultural practices has also decreased the capacity of aerated soils to oxidize CH4 (Ojima et al., 1993; Hütsch, 1998; Smith and Conen, 2004), further enhancing agriculture’s impact on global climate change. Soil carbon sequestration is the increase in soil organic carbon (SOC) corresponding to the net balance between organic C inputs to soil and C losses (Paustian et al., 1997; Lal, 2004; Rees et al., 2005). In irrigated cropping systems, tillage appears to be a main factor influencing CO2 exchanges between soil and atmosphere, but the effects of crop rotation and N fertilization on CO2 exchange are still not clear. The influence of these practices on CH4 fluxes is also not clearly understood. Differences in observed effects are often the result of varying soil and climatic conditions and the duration of management practice adoption. Tillage, N fertilization, and crop rotation influence on CO2 and CH4 flux has been studied in many climates (Ball et al., 1999; Baggs et al., 2006; Omonode et al., 2007), but little information is available from semiarid irrigated cropping systems (Mosier et al., 2006), where dry climate conditions may significantly affect soil microbial processes relative to the more studied humid temperate areas. In particular, the low soil moisture during the noncrop growing season and the different precipitation and irrigation regimes during the cropping season can affect SOC evolution and soil microbial processes. Reduced tillage intensity is proposed as a leading practice in SOC sequestration enhancement through reduced residue and SOM oxiF. Alluvione, Dep of Agronomy, Forest and Land Management, Univ. of Turin, via L. da Vinci 44, 10095 Grugliasco (TO), Italy; A.D. Halvorson and S.J. Del Grosso, USDA–ARS, 2150 Centre Ave., Bldg. D, Suite 100, Fort Collins, CO 80526-8119. Contribution from USDA-ARS, Fort Collins, CO. The U.S. Department of Agriculture offers its programs to all eligible persons regardless of race, color, age, sex, or national origin, and is an equal opportunity employer. Mention of trade names or proprietary products does not indicate endorsement by USDA and does not imply its approval to the exclusion of other products that may be suitable. Abbreviations: CT, conventional moldboard plow tillage; CT-CC, conventional-till continuous corn; DOY, day of year; GHG, greenhouse gas; LSD, least significant difference; MDF, minimum detectable flux; NT, no-till; NT-CB, no-till corn-barley; NT-CC, no-till continuous corn; NT-CDb, no-till corn-dry bean; SOC, soil organic C; SOM, soil organic matter; UAN, urea ammonium nitrate fertilizer; WFPS, water-filled pore space; ρb, soil bulk density.
2023
dation rates (Lal and Kimble, 1997; Follett, 2001; West and Post, 2002). Many studies found decreased CO2 emissions under NT in comparison with CT (Curtin et al., 2000; Chatskikh and Olesen, 2007; Sainju et al., 2008), with very big differences during the high CO2 fluxes that follow tillage operations (Al-Kaisi and Yin, 2005; Omonode et al., 2007; Reicosky and Archer, 2007). In contrast, Franzluebbers et al. (1995) found higher emissions under NT than CT during overnight measurements using alkali traps. Ball et al. (1999) and Omonode et al. (2007) found lack of statistical difference in seasonal CO2 emissions between tillage systems; while different results, varying with year, were found by Mosier et al. (2006) and Fortin et al. (1996). Soil structure disturbance induced by tillage has shown a clear negative effect on soil CH4 oxidation on arable cropland soils when compared to native soils (Mosier et al., 1991; Ojima et al., 1993; Hütsch, 2001), but the ability of soils to recover their CH4 oxidizing capability following adoption of NT management is not well documented. Higher (Hütsch, 1998; Ball et al., 1999; Six et al., 2004) or not different (Jacinthe and Lal, 2005; Mosier et al., 2006; Omonode et al., 2007) oxidizing activity has been found in NT than in CT systems. The time period needed for the recovery appears to be crucial and usually long, in order of decades (Kruse and Iversen, 1995; Hütsch, 2001). Adoption of adequate levels of fertilization has been promoted as an important factor for SOC sequestration through increased residue production (Alvarez, 2005; Follett, 2001; Halvorson et al., 1999) and C addition with organic fertilizers (Grignani et al., 2007), but still unclear is the influence of N addition on CO2 emissions from soil. No affect of mineral N fertilization is commonly found (Rochette and Gregorich, 1998; Amos et al., 2005; Mosier et al., 2006), but in some cases N fertilization tends to either decrease (Al-Kaisi et al., 2008; Wilson and Al-Kaisi, 2008) or increase (Sainju et al., 2008) CO2 emissions. Also unclear are the consequences of N fertilization on CH4 fluxes. Nitrogen addition is known to strongly reduce CH4 oxidation in forest and grassland soils as a consequence of biochemical competition, while in arable cropland soils the effect can be null or of reduction (Hütsch et al., 1993; Hütsch, 2001; Mosier et al., 2006). Few studies have analyzed cropping system effects on CO2 or CH4 soil-atmosphere exchanges. Crop rotation can influence SOC sequestration through the amount and quality of residue returned to soil (Curtin et al., 2000; Omonode et al., 2007; Wilson and Al-Kaisi, 2008), but this is not always reflected in different soil respiration rates (Sainju et al., 2008). In nonflooded agricultural ecosystems, a particular crop generally does not directly influence soil CH4 oxidation (Robertson et al., 2000; Mosier et al., 2006), but it is more likely influenced by secondary factors linked to crop management, such as different agrochemicals used or soil compaction (Hütsch, 2001). We evaluated (i) the potential of no-till to reduce CO2 emissions and to recover the soil CH4 oxidizing ability relative to the conventional moldboard plowing, (ii) the possible effect of mineral N fertilization on CO2 fluxes, through different amount of crop residue returned to the soil, and the ability of reduced N rates to increase soil CH4 oxidation, and (iii) the effect of 2024
crop rotation on CO2 fluxes, through the addition of different amount of residue with different quality, and on net CH4 flux. We continued the work of Mosier et al. (2006) on the very same plots with the objective of following the evolution of the effect of adopted practices on greenhouse gas (GHG) emissions, as their influence is known to be time dependent and results may vary during the transition period before an equilibrium is reached (Hütsch, 2001; Six et al., 2004). The evaluation of the intermediate effect is fundamental to better understand the processes involved, the time needed before a result is achieved, and because some practices can be adopted by farmers only for a limited period and then changed, thus the temporary effect needs to be known. Carbon dioxide and CH4 fluxes reported are contemporary to the N2O fluxes presented by Halvorson et al. (2008).
Materials and Methods Site Description and Experimental Treatments The research site was located at the Agricultural Research Development and Education Center (ARDEC) in northeastern Colorado near Fort Collins (40°39´ N, 104° 59´ W; 1535 m a.s.l.). The region has a semiarid temperate climate with mean annual temperature of 10.6°C and mean annual precipitation of 383 mm (average from 1900–2005). Corn (Zea mays L.), winter wheat (Triticum aestivum L.), barley (Hordeum distichon L.), and dry bean (Phaseolus vulgaris L.) are the main crops grown in the region. Soil at the site is a Fort Collins clay loam (fine-loamy, mixed, superactive, mesic Aridic Haplustalf ), with an average SOC content of 12 g kg–1, average C/N ratio of 8.1, and a pH of 7.8. Selected soil chemical and physical properties at the research site for the different treatments are reported in Halvorson et al. (2008) and in Zobeck et al. (2008). The field site was cropped beginning in about the 1950s and was managed under CT continuous corn for 6 yr before initiating the experiment in 1999. Four irrigated corn-based cropping systems were compared during the 2005 and 2006 growing seasons: continuous corn under conventional moldboard plow tillage (CT-CC), continuous corn under no-till (NT-CC), NT corn-barley (NT-CB), and NT corn-dry bean (NT-CDb), each under different levels of N fertilization. Mechanical tillage was used in the CT-CC plots, with the main tillage operations (disk and moldboard plow to 25 cm depth) performed in fall after corn harvest and with another disk operation followed by a roller-mulcher (two passes) and a land leveler (two passes) during seedbed preparation in spring before planting. Crop residues were left on the soil surface of the NT plots after harvest. Four N rates [0 (N1), 67 (N2), 134 (N3), and 246 (N4) kg N ha–1] were monitored both years in the CT-CC and NT-CC systems. In the other cropping systems, the zero N (N1) and high N (N4) treatments were monitored. The high N rate was 246 kg N ha–1 for corn, 56 kg N ha–1 for dry bean, and 112 kg N ha–1 for barley. In 2005, urea-ammonium nitrate (UAN) was preplant, subsurface-band applied parallel to crop rows (33 cm band spacing) at the full N rate to the NT-CB (29 March, day of year [DOY] 88) and NT-CDb (24 May, DOY 144) rotations; however, 50% of the N rate was applied as UAN to the CT-CC and NT-CC systems on 25 April (DOY 115). The other
Journal of Environmental Quality • Volume 38 • September–October 2009
50% of the N rate was broadcast applied as a polymer-coated urea (ESN, registered trade mark of Agrium Advanced Technologies, Sylacauga, AL) on 9 June (DOY 160) to the CT-CC and NT-CC systems in 2005. A split N application was used in 2006 for all cropping systems, with 50% of the N rate surface band applied near the corn row as ESN on 17 May (DOY 137), and the other 50% of the N rate surface band applied near the corn row as dry granular urea on 12 June (DOY 163) followed by 38 mm irrigation water on 13 June (DOY 164). The experimental design for each cropping system was a randomized complete block with three replications. Further information about the study was reported by Halvorson et al. (2008). Herbicides were used for weed control in all treatments, resulting in the plots being relatively weed-free. Only one herbicide that may have affected the CH4 oxidizing ability of the soil (Arif et al., 1996), 2,4-dichlorophenoxyacetic acid (commonly referred to as 2,4-D), was applied during the 2005 growing season to the NT-CB plots (DOY 124; 0.184 kg a.i. ha–1) and before corn planting (DOY 110; 0.092 kg a.i. ha–1) in 2006 to the NT crop rotations. The maximum and minimum daily air temperatures, irrigation, and precipitation events for the growing season were reported in detail by Halvorson et al. (2008). Briefly, weather conditions varied slightly during the growing season between years. Air temperatures from May through August averaged 17.6°C in 2005 and 17.7°C in 2006. Precipitation was 102 mm higher in 2005 (177 mm from DOY 124–279) than in 2006 (65 mm from DOY 122–275). A lateral-move sprinkler irrigation system was used to apply irrigation water as needed during the growing season [determined weekly by the feel method (Klocke and Fischbach, 1998)]. Irrigation amount varied with crop grown in 2005 (NT-CB, 184 mm; NT-CC and CT-CC, 388 mm after corn planting; NT-CDb, 299 mm), but was the same for all rotations in 2006 as all were planted to corn (403 mm, of which 377 mm applied after corn planting). In 2005, 25 mm more irrigation water was applied to CT-CC plots than the NT-CC plot before corn planting to ensure germination of the corn due to very dry soil conditions. In 2006, irrigation was not performed from 18 July (DOY 199) to 1 August (DOY 213) because of a mechanical problem with the irrigation system. Coupled with average daily maximum air temperatures of 33°C, this problem resulted in severe water stress during corn pollination and kernel set contributing to reduced corn yields in 2006 (Halvorson, unpublished data, 2008). Additional plot management details for the study are provided in Halvorson et al. (2008).
Gaseous Fluxes and Ancillary Measurements Measurement of the soil-atmosphere exchange of CO2 and CH4 began about 3 d after crop planting and were performed during the 2005 and 2006 crop growing seasons (in 2005: from DOY 124–279 in NT-CC and CT-CC, from DOY 146–304 in NT-CDb, from DOY 90–243 in NT-CB; in 2006: from DOY 122–275 in all rotations) following the same procedures reported by Mosier et al. (2006). Net CH4 emissions (the combined effect of positive emissions into and negative consumption from the atmosphere) were measured. Briefly, fluxes were monitored one to
three times per week, mid-morning of each sampling day. A vented nonsteady state closed chamber technique was used (Livingston and Hutchinson, 1995). A rectangular aluminum chamber (78.6 by 39.3 by 10 cm height) with a sampling port was placed in a water channel welded onto an anchor that had been inserted 10 cm into the soil at each sampling site. Chamber tops were insulated and covered with a light-reflective sheet to limit temperature increases inside the chamber (Hutchinson and Livingston, 2002; Rochette and Eriksen-Hamel, 2008). Possible small soil temperature increases were considered uninfluential on microbial processes during the short 30 min deployment time (Hutchinson and Livingston, 1993; Rochette and Hutchinson, 2005). Anchors were set perpendicular to the crop row so the crop row and interrow were contained within each chamber. Anchors were installed each year 3 d before beginning measurements and were not removed until fall. Duplicate flux measurements were made within each plot for a total of six measurements per treatment. The plants were cut off when they became too tall to fit inside the chambers. Gas samples from inside the chambers were collected by syringe at 0, 15, and 30 min after chamber closure. Twenty five milliliter air samples were withdrawn from chamber head space and injected into 12 mL evacuated tubes that were sealed with butyl rubber septa (Exetainer vial from Labco Limited). Carbon dioxide and CH4 concentrations were quantified using a fully automated Varian 3800 gas-chromatograph equipped with thermal conductivity and flame ionization detectors (Mosier et al., 2005). Fluxes were calculated from the linear or nonlinear (Hutchinson and Mosier, 1981) increase in concentration (selected according to the emission pattern) in the chamber headspace with time as suggested by Livingston and Hutchinson (1995). Minimum detectable flux (MDF) for CO2 and CH4 calculated according to Chan et al. (1998) taking into account the analytical and sampling error corresponded to 256.0 g CO2–C ha–1 d–1 and 1.0 g CH4–C ha–1 d–1, respectively. Estimates of daily CO2 and CH4 flux between sampling days were made using a linear interpolation between adjacent sampling dates. As Chan and Parkin (2001) did for average values, we followed the recommendation of Gilbert (1987) and included daily fluxes that could not be considered statistically different from 0, because they were below the minimum detectable flux, in the cumulative value calculation. Volumetric soil water content (0–10 cm depth) and soil (5 cm depth) and air temperature were monitored during each gas sampling event in all plots using soil dielectric constant probes (ECHO probes from Decagon Devices, Inc.) and temperature probes (HH21 Digital Thermometer from Omega Engineering, Inc.), respectively. Water filled pore space (WFPS) was calculated according to the soil bulk density (ρb) measured at 0- to 7.6-cm depth in each plot following crop harvest and a particle density of 2.65 Mg m–3 (Linn and Doran, 1984).
Data Analysis Tillage, crop rotation, and N effects on CO2 and CH4 flux, soil temperature, WFPS during the growing season, and ρb in fall were analyzed. To better understand the treatments effects on net CH4 flux, the percentage of days in which CH4 flux was negative and the net CH4 oxidation (negative fluxes) and emission
Alluvione et al.: Carbon Dioxide and Methane Fluxes from Irrigated Cropping Systems
2025
(positive fluxes) rates were analyzed as well (Chan and Parkin, 2001). Two comparisons were made: tillage and N effects were evaluated comparing CT-CC and NT-CC corn treatments in 2005 and 2006, while crop rotation and N effects were evaluated comparing NT-CB, NT-CC, and NT-CDb separately in 2005 and 2006. Crop effects were analyzed by comparing NT-CC, NT-CB, and NT-CDb in 2005 when corn, barley, or dry bean, respectively, was grown depending on the rotation. The NT-CC, NT-CB, and NT-CDb rotations were compared in 2006 at the 0 and 246 kg N ha–1 to provide information on previous crop and N effects on CO2 and CH4 flux in corn. Cumulative CO2 and CH4 flux during the 2005 and 2006 growing seasons gave information on the effects of crop rotation. However, since the length of the growing season was different for different crops, a daily average was used for this last comparison. Differences in CO2 and CH4 daily and cumulative fluxes, soil temperature, WFPS, and ρb by tillage, N rate, crop rotation, year, and their interactions were determined by ANOVA using Analytical Software Statistix8 program (Analytical Software, Tallahassee, FL). When significantly different (P < 0.05), means were separated using the least significant difference (LSD) method. Regressions between CO2 and CH4 daily fluxes and soil temperature and WFPS were analyzed using SPSS 16.0 (SPSS Inc., Chicago, IL).
Results and Discussion The comparisons made between CT-CC and NT-CC or among NT-CB, NT-CC, and NT-CDb indicated no interactions between tillage and N fertilization or crop rotation and N fertilization effects on CO2 and CH4 flux (Table 1). Consequently, individual treatment effects are reported separately.
Tillage Effect Soil temperatures at planting tended to be 1 to 2°C warmer in CT soils compared to NT soils with this trend reversing during the latter part of the growing season (Fig. 1 and 2). As reported by Halvorson et al. (2008), soil temperature was significantly cooler in the NT-CC than in the CT-CC during May and warmer during September and October both in 2005 and 2006. Similar seasonal trends in temperature were also found by Fortin et al. (1996), Mosier et al. (2006), and Wagai et al. (1998). When averaged across all months, growing season soil temperatures were not significantly different between NT-CC and CT-CC systems in both years (Halvorson et al., 2008). Average WFPS values were always higher in NT-CC (2005: 55.2%; 2006: 51.6%) compared to CT-CC (2005: 47.9%; 2006: 50.4%), but significantly different only in 2005 (Table 1). Similar to our results, Kessavalou et al. (1998) found higher WFPS in NT than in CT. Soil bulk density measured in fall was significantly (Table 1) lower in CT-CC (1.33 Mg m–3) than in NT-CC (1.40 Mg m–3). No differences in ρb between years were found. Carbon dioxide emissions during the growing season followed the common parabolic pattern that tracks the combined effect of residue decomposition and root activity (Rochette and Flanagan, 1997; Rochette et al., 1999; Amos et al., 2005) with
2026
trend toward increasing fluxes during the first part of the growing season (May and June) and decreasing fluxes in the latter part (from July–September) (Fig. 1 and 2). Tillage management significantly influenced the intensity of emissions through both seasons. In 2005, NT-CC significantly decreased CO2 flux compared to CT-CC until the end on July, while during the remainder of the growing season emissions were either greater under NT-CC or not different between tillage systems (Fig. 1). In 2006, NT-CC showed constantly lower CO2 emissions than CT-CC through the growing season, even when differences between tillage treatments became smaller during the second half of August when soil temperatures tended to be higher in NTCC (Fig. 2). These patterns resulted in a significant tillage by year interaction with respect to cumulative CO2 fluxes (Table 1). Lower CO2 cumulative emissions were found in NT-CC (2005: 2.17 Mg CO2–C ha–1; 2006: 1.98 Mg CO2–C ha–1) than in CTCC (2005: 2.23 Mg CO2–C ha–1; 2006: 2.57 Mg CO2–C ha–1), with this difference being significant only in 2006. Averaged over the 2 yr, CO2 cumulative emissions during the growing season were 14% lower in NT-CC (2.08 Mg CO2–C ha–1) than in CTCC (2.41 Mg CO2–C ha–1). Data in 2005 and 2006 confirmed the lower CO2 emissions in NT-CC than CT-CC as earlier observed by Mosier et al. (2006) on the very same plots. However, cumulative emissions were higher as the whole year was monitored (winter fallow and tillage operations included), while in our study only the corn growing season fluxes were measured. Similar emission intensities were observed by Omonode et al. (2007), even though no tillage effect was found. Soil organic C evolution is the result of the combined effect of C inputs to and C losses from the soil (Paustian et al., 1997; Lal, 2004; Rees et al., 2005). Similar to Fortin et al. (1996), we evaluated if the difference in cumulative CO2 emissions between tillage treatments was a viable estimate of the SOC change with time. In this case, it is important that C inputs to soil are similar, while CO2 emissions reflect differential C losses. In the treatments under evaluation, C inputs are represented only by aboveground crop residues. Considering that the amount and C/N ratio of the aerial biomass in CT and NT were not significantly different (Halvorson, unpublished data, 2008), SOC evolution should not have been determined by this factor. With regard to CO2–C losses, the authors speculate that similar aboveground biomass development contributed to similar amounts of CO2 produced by roots and root exudates respiration under CT-CC and NT-CC. Thus, the differences between treatments were probably mainly due to different rates of crop residue and SOM oxidation. Similarly, Fortin et al. (1996) assumed that differences in CO2 emissions between CT and NT were reflecting differences in the rate of change of SOC as the amount of residue produced in the two treatments was similar. Moreover, Rochette et al. (1999) effectively found lower percentage of corn residue C losses with NT (35%) than CT (40%). Basing our calculation on this assumption, the difference in CO2–C losses during the corn growing season between CT-CC and NT-CC was 331 kg CO2–C ha–1 yr–1. As the emissions during the noncrop growing seasons were not measured in this study, we estimated
Journal of Environmental Quality • Volume 38 • September–October 2009
Table 1. Probability (P) values for the ANOVA analyses. Comparison CT-CC vs. NT-CC
Soil temp. WFPS ρb CO2 CH4 CH4 CH4 CH4
NT-CB vs. NT-CC vs. Soil temp. NT-CDb Soil temp. WFPS WFPS ρb ρb CO2 CO2 CO2 CH4 CH4 CH4 CH4 CH4 CH4 CH4 CH4 CH4
Variable average (2005 and 2006) average (2005 and 2006) fall sampling (2005 and 2006) cumulative flux (2005 and 2006) cumulative flux (2005 and 2006) percentage of negative days (2005 and 2006) average net oxidation rate (2005 and 2006) average net emission rate (2005 and 2006) average (2005) average (2006) average (2005) average (2006) fall sampling (2005) fall sampling (2006) cumulative flux (2005) cumulative flux (2006) average daily flux (2005 and 2006) cumulative flux (2005) cumulative flux (2006) average daily flux (2005 and 2006) percentage of negative days (2005) percentage of negative days (2006) average net oxidation rate (2005) average net oxidation rate (2006) average net emission rate (2005) average net emission rate (2006)
T NS‡