- Top of page
- Materials and methods
Around 4.4 million ha of land in USDA Conservation Reserve Program (CRP) contracts will expire between 2013 and 2018 and some will likely return to crop production. No-till (NT) management offers the potential to reduce the global warming costs of CO2, CH4, and N2O emissions during CRP conversion, but to date there have been no CRP conversion tillage comparisons. In 2009, we converted portions of three 9–21 ha CRP fields in Michigan to conventional tillage (CT) or NT soybean production and reserved a fourth field for reference. Both CO2 and N2O fluxes increased following herbicide application in all converted fields, but in the CT treatment substantial and immediate N2O and CO2 fluxes occurred after tillage. For the initial 201-day conversion period, average daily N2O fluxes (g N2O-N ha−1 d−1) were significantly different in the order: CT (47.5 ± 6.31, n = 6) ≫ NT (16.7 ± 2.45, n = 6) ≫ reference (2.51 ± 0.73, n = 4). Similarly, soil CO2 fluxes in CT were 1.2 times those in NT and 3.1 times those in the unconverted CRP reference field. All treatments were minor sinks for CH4 (−0.69 ± 0.42 to −1.86 ± 0.37 g CH4–C ha−1 d−1) with no significant differences among treatments. The positive global warming impact (GWI) of converted soybean fields under both CT (11.5 Mg CO2e ha−1) and NT (2.87 Mg CO2e ha−1) was in contrast to the negative GWI of the unconverted reference field (−3.5 Mg CO2e ha−1) with on-going greenhouse gas (GHG) mitigation. N2O contributed 39.3% and 55.0% of the GWI under CT and NT systems with the remainder contributed by CO2 (60.7% and 45.0%, respectively). Including foregone mitigation, we conclude that NT management can reduce GHG costs by ~60% compared to CT during initial CRP conversion.
- Top of page
- Materials and methods
The USDA Conservation Reserve Program (CRP) builds contracts with agricultural landowners in the United States to retire highly erodible and environmentally sensitive cropland and pasture into perennial vegetation for periods ≥10 years. The program, established by the Food Security Act of 1985, is designed to reduce soil erosion, improve water and air quality, enhance wildlife populations, and to sequester carbon in soil and biomass. In 2007, as many as ~15 million ha were enrolled, representing ~9% of total US cropland (Economic Research Service (ERS), 2011; Farm Service Agency (FSA), 2012). Since then, enrolled land had decreased to ~12 million ha in 2012, and an additional ~4.4 million ha of land are in CRP contracts that will expire between 2013 and 2018 (Farm Service Agency (FSA), 2012). Higher prices for corn (Zea mays L.) and other crops and expanded biofuel production are expected to induce farmers to return CRP land to grain production (Du et al., 2008; Secchi et al., 2009). Many environmental benefits may subsequently be lost. Of particular concern are changes to greenhouse gas (GHG) emissions – fluxes of CO2, nitrous oxide (N2O) and methane (CH4) during and after conversion (CAST (Council for Agricultural Science & Technology), 2011).
Grassland conversion into crop production can accelerate both soil C and nitrogen (N) cycles, and results in significant GHG emissions. In particular, land conversion practices such as plowing can enhance soil organic matter oxidation, nitrification, and denitrification and substantially increase CO2 and N2O emissions (Pinto et al., 2004; Grandy & Robertson, 2006a; Nikièma et al., 2012). No-till (NT) offers the potential to attenuate such increases, but to date there have been no GHG comparisons of NT and conventional tillage (CT) during CRP conversion.
The effects of tillage on soil carbon are well known. Plowing mixes crop residues with the soil, increases the aeration of surface soil and reduces soil aggregation, all of which enhances organic matter decomposition and CO2 release (Haas et al., 1957; Buyanovsky & Wagner, 1998; Grandy & Robertson, 2006b; Regina & Alakukku, 2010). In contrast, the soil under NT is left undisturbed. More stable aggregates under NT protect soil organic carbon (SOC) from microbial decomposition and allow SOC storage (Six et al., 2000). Dolan et al. (2006), for example, reported that NT managed soil contained over 30% more SOC than CT soils to 20 cm after 23 years of NT. Syswerda et al. (2011) reported ~11% higher SOC to 1 m depth under NT than CT after 12 years of NT. West & Post (2002) used a global database of 67 long-term agricultural experiments to estimate that conversion from CT to NT can annually sequester 48 ± 13 g C m−2 yr−1 in surface horizons. There is little evidence for statistically different changes at deeper depths (Kravchenko & Robertson, 2011). Following CRP conversion, Follett et al. (2009) reported no SOC change (0–30 cm depth) within 6.5 years after conversion of CRP grasslands to NT corn in Nebraska. Anken et al. (2004), however, reported that SOC (0–20 cm depth) decreased under both NT and CT similarly in Switzerland for the first 7 years after conversion of a 10 year old grassland to maize-winter wheat production.
Effects of CT on soil N2O emissions compared to NT are still in debate. Agricultural soil N2O emissions account for about 60% of global total anthropogenic N2O production (IPCC (Intergovernmental Panel on Climate Change), 2007) due to two microbial processes: denitrification and nitrification (Robertson & Groffman, 2007). Theoretically, NT can strongly affect both these processes through effects on soil water, carbon, pore space, and soil N concentrations. In practice, some studies have shown higher N2O emissions from NT than CT (e.g., Baggs et al., 2003; Rochette et al., 2008), with higher rates in NT mostly attributed to restricted soil aeration due to higher water content, which is conducive to denitrification. However, others have found lower emissions in NT than CT, attributed to improved soil structure and lower soil temperatures (e.g., Chatskikh & Olesen, 2007; Ussiri et al., 2009). Still others have found no difference between NT and CT (e.g., Robertson et al., 2000; Choudhary et al., 2002; Boeckx et al., 2011).
Methane oxidation is also affected by agricultural management. CH4 oxidation by methanotrophic bacteria in well-aerated soils is an important sink (5%, globally) for atmospheric CH4 IPCC (Intergovernmental Panel on Climate Change), 2007. In theory, a less disturbed soil structure and improved gas diffusion in NT should enhance the CH4 oxidation capacity of methanotrophic bacteria relative to CT (Six et al., 2004; Ussiri et al., 2009). However, studies to date have reported no significant NT effects on CH4 oxidation rates (Robertson et al., 2000; Jacinthe & Lal, 2005).
In an earlier study, Gelfand et al. (2011) reported that the conversion of CRP land to NT soybean production released significant amounts of CO2 and N2O and had little effect on CH4 oxidation rates. Here, we extend their results to examine the impact of CT practices on GHG fluxes during conversion. Specifically, we hypothesize that for the CRP conversion year, NT relative to CT will (i) attenuate N2O emissions; (ii) reduce C loss; and (iii) avoid the loss of CH4 oxidation. Furthermore, we evaluate the relative importance of each flux to the overall GHG cost of CRP conversion.
- Top of page
- Materials and methods
The conversion of our CRP grasslands into row crops resulted in a substantial GHG release that differed by tillage practice. The most remarkable difference between CT and NT management during conversion was in N2O fluxes. We found immediate and substantial tillage-induced N2O emissions under CT that exceeded the CO2-equivalent loss of soil C over the 201-day study period. Total N2O emissions under converted CT soybean were 2.1-fold higher than under converted NT soybean and 18.8-fold higher than under unconverted smooth brome grass (reference field). The magnitude of CT N2O emissions exceeded that of fertilizer-induced N2O fluxes in the same area (Robertson et al., 2000; Hoben et al., 2011). Even with NT practices, however, CRP conversion still caused large N2O emissions, with fluxes under NT 5.3 times higher than under unconverted reference.
Soil CO2 emissions under CT were also significantly higher than those under NT and reference treatments. Cumulative NEE of CO2 under CT were 2.2-fold higher than those under NT over the study period. The converted fields under both CT and CT were carbon sources under both CT and NT, whereas the unconverted reference treatment was a net carbon sink. All treatments were a small sink for atmospheric CH4. However, changes in CH4 oxidation rates did not contribute significantly to the GWI of conversion compared with N2O and CO2. Overall, N2O accounted for 39.3% of the net GWI of conversion under CT and 55.0% under NT with the remainder contributed by CO2 (60.7% and 45.0%, respectively), excluding the CO2 costs of herbicide and fuel, which were negligible (Gelfand et al., 2011).
Nitrous oxide (N2O) fluxes increased 18- to 55-fold immediately on the first day after tillage operations in all CT treatments. Over the study period, mean daily CT N2O emissions (47.5 ± 6.3 g N2O-N ha−1 d−1) were relatively higher than those reported for fertilized annual crops at a nearby site (3.35 ± 0.30 g N2O-N ha−1 d−1) (Robertson et al., 2000) and for heavily fertilized crops elsewhere in Michigan (25.8 g N2O-N ha−1 d−1from corn fertilized at 225 kg N ha−1) (Hoben et al., 2011). Similar substantial amounts of N2O emissions following tillage have been reported for other studies where unmanaged vegetation has been converted to cropland. For example, Grandy & Robertson (2006a) reported a 3.1 to 7.7 fold increase in N2O emissions after plowing long-term undisturbed grassland over a 3 year period. Nikièma et al. (2012) reported high N2O fluxes of 57.2 and 41.8 g N2O-N ha−1 d−1 after converting heavily manured pastureland (200 kg N ha−1 yr−1) to poplar and willow production, respectively. Possible reasons for high N2O emissions could be increased production of available N and C after SOM mineralization (Grandy & Robertson, 2006a) and increased substrate supply to nitrification and denitrification after the incorporation of residues into the soil (Piva et al., 2012). In contrast, daily N2O fluxes under NT also continuously increased from 1.93 ± 0.75 to 66.7 ± 16.0 g N2O-N ha−1 d−1 for the first 45 days after herbicide application, but overall rates were approximately one third of those from under CT. Available C and N from decomposed dead grass and roots are likely reasons.
In the unfertilized fields studied here, available N could be one of the most important driving factors for N2O emissions. Accelerated N mineralization from SOM and incorporated residue after tillage can increase available N and thus enhance nitrification and denitrification. Resin strip measurements indicate that for the 37 day period after CT tillage, soil NO3−-N and NH4+-N concentrations under CT (57.7 ± 7.16 and 2.30 ± 0.41 μg N cm−2) were substantially higher than those under NT (12.1 ± 1.59 and 0.31 ± 0.08 μg N cm−2), respectively. Daily N2O fluxes were strongly correlated with total available N from resin strip measurements (N2O fluxes = 34.8 × EXP (0.36 × available N), R2 = 0.19, P < 0.01). However, NO3−-N and NH4+-N concentrations in soil cores showed no consistent differences between CT and NT. This is likely because soil-KCl extractions measure only the soil available N pool size. This pool can be rapidly utilized by microbes and plants or leached out of the soil so that it cannot be detected accurately, especially when the N pool size is small. In contrast, ion exchange strips measure both the soil available N pool and the flux of N ions through the mineral pool (Bowatte et al., 2008). In this study, the resin strips provided the more interpretable results.
Soil N2O fluxes were also likely affected by available soil carbon (Dalal et al., 2003; Wang et al., 2011). Firstly, killed and incorporated brome grass, in conjunction with dead roots, provided heterotrophic denitrifiers with more available carbon and as well will have increased O2 demand. CO2 as an end product of decomposition indicated the extent to which dead brome grass was decomposed. Especially during the period between herbicide application (May 5) and tillage operations in the CT treatment (June 8), soil CO2 emissions were 5.7 times, those of emissions from the unconverted reference field, indicating that more decomposition took place in the herbicide applied fields than in the reference field. In addition, the old CRP land had accumulated relatively high amounts of SOC, which has a potential to provide more available carbon for N2O production due to SOC decomposition after tillage. Compared to SOC at nearby LTER experimental sites (Syswerda et al., 2011), SOC concentrations in our studied fields (21.3 ± 0.8 g C kg−1 soil) prior to the conversion were comparatively higher than annual crops under CT (10. 4 ± 3.4 g C kg−1 soil) and NT (11.5 ± 0.4 g C kg−1 soil) and close to deciduous forest levels (24.0 ± 3.4 g C kg−1) for 0–20 cm depth. In addition, enhanced SOM decomposition will consume oxygen and create localized anaerobic conditions favoring denitrification (Wang et al., 2011).
Soil N2O fluxes are also affected by soil water content. Two relatively larger N2O peaks occurred after rainfall in this study when WFPS% was >60%. The possible reason is that rainfall events create anaerobic conditions, which can stimulate N2O emissions from denitrification. This finding has been reported by many studies (e.g., Elder & Lal, 2008; Wang et al., 2011). However, in this study overall N2O fluxes showed no significant correlation with soil moisture (P > 0.05). Wet soil conditions did not necessarily give rise to high N2O emission. For example, soil N2O fluxes in the reference field remained low and stable through the whole study period even after considerable rainfall. In addition, we observed low emissions of N2O at all fields after September even when WFPS% was larger than 60% following rainfall. For both cases, this indicates that N2O production was likely restricted by other more limiting factors such as available N or low temperature.
The comparison between NT and CT N2O fluxes has been widely studied and it is still difficult to generalize. Six et al. (2004) analyzed 44 comparisons of N2O emissions under CT and NT globally and found higher N2O emissions in the first 10 years of NT than CT and thereafter similar or lower N2O emissions under NT. They argued that increased soil water content under NT promoted denitrification and thus enhanced N2O production in the first 10 years. A more recent study using a meta-analysis of 239 direct comparisons between CT and NT/reduced tillage (Van Kessel et al., 2013) found no N2O emission differences. However, in this study, CRP land with its long-term no-till history and high SOC may provide a special case. Our results suggest that adopting NT practices can significantly reduce N2O emissions compared to CT, but NT management cannot eliminate the cost of N2O emissions during CRP conversion.
Soil CO2 emissions under both CT and NT soybeans were significantly higher than those in unconverted reference fields (P < 0.05). Two possible reasons are (i) decomposition of dead grass and roots in the soil; and (ii) accelerated SOM decomposition after tillage. In addition, soil CO2 emissions in CT soybean were higher than emissions in NT soybean (P < 0.05). Similar results have been reported in many studies (e.g., Grandy & Robertson, 2006a; Chatskikh & Olesen, 2007; Alluvione et al., 2009). Tillage enhanced SOC decomposition and thus increased CO2 release to the atmosphere.
Soil CO2 fluxes can be governed by soil temperature, moisture, and other factors. Multiple Linear regressions of soil CO2 fluxes with soil temperature and WFPS% showed no significant correlation between CO2 fluxes and WFPS%, although WFPS% might have affected CO2 emission at some specific times during the drought period in July with its relatively low emissions. On the other hand, a positive relationship was found between soil CO2 fluxes and soil temperature: soil CO2 fluxes = 11.5 × EXP (0.07 × soil temperature), R2 = 0.21, P < 0.01). Exponentially increased soil CO2 fluxes with rising temperature have been reported by many studies (e.g., Lloyd & Taylor, 1994; Reichstein & Beer, 2008; Almaraz et al., 2009).
The NEE of CO2 fluxes for CT soybeans was more than twice that for NT soybeans, and the converted fields under both CT and NT were net sources for CO2. This is because carbon released from the decomposition of grass residue and SOC exceeded the carbon uptake from photosynthesis in converted fields. On the contrary, the unconverted reference field was a net sink for atmospheric CO2.
The range of daily CH4 fluxes (−6.4 to 4.5 g CH4-C ha−1 d−1) we observed were similar to CH4 fluxes of −1.80 ± 0.06 g CH4-C ha−1 d−1 for cropland in Michigan (Robertson et al., 2000). All fields were net sinks for CH4, although some other studies found cropland under CT could be a small net source (Alluvione et al., 2009; Ussiri et al., 2009). Fluxes in CO2 equivalents were negligible compared with CO2 and N2O fluxes, which had generally been reported for other upland cropping systems (Robertson et al., 2000; Wang et al., 2011).
No statistically significant differences in CH4 oxidation rates were found among any treatments, although oxidation rates in CT were 62.9% and 38.8% lower than those in the NT and reference treatments, respectively. Similar results of no differences between CT and NT systems have been reported in some studies for sites nearby (Robertson et al., 2000; Suwanwaree & Robertson, 2005). However, other studies reported higher oxidation rates in NT than CT or uptake in NT but net emissions in CT (Ussiri et al., 2009). They attributed this to undisturbed soil structure and greater gas diffusion under NT. Another possible reason was that increased mineralization after tillage enhanced NH4+ production, and NH4+ could competitively inhibited CH4 oxidation. In addition, we found no significant difference in CH4 oxidation before and after conversion of CRP land, although some studies have found that the CH4 oxidation rates of a grassland were reduced by 75% after only 8 months of conversion into CT cropland (Ball et al., 1999) or higher CH4 oxidation rates in midsuccessional grassland than cropland (Robertson et al., 2000). It seems likely that CH4 oxidation rates had not increased under 20 years of CRP brome grass sufficiently to be significantly re-suppressed by cropping.
Methane (CH4) oxidation rates can also be regulated by soil water content and soil temperature. CH4 oxidation rates were found negatively correlated with soil water content in some studies, probably due to limited CH4 diffusion in the wet soil (Del Grosso et al., 2000; Khalil & Baggs, 2005). However, CH4 oxidation may be inhibited in dry soils (Khalil & Baggs, 2005). In this study, no apparent seasonal CH4 flux patterns were observed. We found CH4 fluxes were not significantly related with either WFPS% or soil temperature in any treatments, although other studies have shown CH4 flux from NT to be negatively correlated with soil temperature (Ussiri et al., 2009).
Global warming impact
Over the study period (201 days), the GWI of converted soybean fields was 11.5 and 2.87 Mg CO2e ha−1 for CT and NT operations, respectively, whereas the GWI of the unconverted CRP reference field was −3.5 Mg CO2e ha−1(Fig. 6). The positive GWI of the converted fields indicates net GHG emissions to the atmosphere, while the negative GWI in the reference field indicates on-going GHG mitigation. The possibility that increased N2O emissions might offset the enhanced soil carbon sequestration in NT systems has been a concern for adopting NT practices (Six et al., 2002; Li et al., 2005), but this was not the case in this study. NT played an important role in reducing GWI compared to CT, by significantly decreasing N2O emissions and reducing SOC loss.
The CT system exhibited a net positive GWI of 11.5 Mg CO2e ha−1. In this system, about 39.3% of the GWI was contributed by N2O production (4.52 Mg CO2e ha−1) even in the absence of synthetic N fertilizer additions. SOC loss as indicated by net CO2 emissions contributed the remainder (60.7% or 6.98 Mg CO2e ha−1). For the NT system, net GWI was 2.87 Mg CO2e ha−1, about 55.0% of which was contributed by N2O production (1.57 Mg CO2e ha−1) with the remaining 45% from CO2 emissions (1.30 Mg CO2e ha−1). The contribution of CH4 oxidation was negligible (<0.1%) under both CT and NT systems.
In contrast to converted fields, the unconverted reference fields showed a net mitigation potential of −3.50 Mg CO2e ha−1 due to very low rates of N2O production and a net uptake of CO2.
The net mitigation potential for the unconverted reference fields indicates that the conversion of CRP land not only increases the emissions of GHGs but also causes the loss of the CRP land's net GHG mitigation ability: 3.5 Mg CO2e ha−1 mitigation would have happened had no conversion occurred. This foregone mitigation capacity must be added to the post conversion GHG fluxes to provide a total net GWI (Gelfand et al., 2011). This yields a total initial cost of 6.4 Mg CO2e ha−1 for NT and 15.0 Mg CO2e ha−1 for CT soybean. Thus, NT can reduce GHG costs by ~60% as compared to CT.
Robertson et al. (2000) calculated for a nearby site under the same soil series that NT practices sequestered 30 g C m−2 yr−1. Based on this rate, CRP conversion by CT rather than NT cost ~8 years of NT carbon sequestration with a single tillage event.
Over time, this additional cost will change depending on future management. If planted with perennial biofuel crops (no tillage and no N fertilization), the plowed soils will stop losing and begin re-accumulating soil carbon and N2O fluxes will likely be low. In contrast, if planted with annual grain crops that are plowed and fertilized every year, soil carbon will continue to be lost until the soil equilibrates (to ~10.4 g C kg−1 soil from annual crops under CT at the nearby KBS LTER site). N2O production differences due to CT and NT will likely diminish (Van Kessel et al., 2013) but N2O fluxes will continue to be high due to N fertilization (Hoben et al., 2011).