Integrated soil fertility management drives the effect of cover crops on GHG 1 emissions in an irrigated field 2

9 Agronomical and environmental benefits are associated with replacing winter fallow by 10 cover crops (CC). Yet, the effect of this practice on nitrous oxide (N2O) emissions 11 remains poorly understood. In this context, a field experiment was carried out under 12 Mediterranean conditions to evaluate the effect of replacing the traditional winter fallow 13 (F) by vetch (Vicia sativa L.; V) or barley (Hordeum vulgare L.; B) on greenhouse gas 14 (GHG) emissions during the intercrop and the maize (Zea mays L.) cropping period. 15 The maize was fertilized following Integrated Soil Fertility management (ISFM) 16 criteria. Maize nitrogen (N) uptake, soil mineral N concentrations, soil temperature and 17 moisture, dissolved organic carbon (DOC) and GHG fluxes were measured during the 18 experiment. The ISFM resulted in low cumulative N2O emissions (0.57 to 0.75 kg N2O19 N ha), yield-scaled N2O emissions (3-6 g N2O-N kg aboveground N uptake) and N 20 surplus (31 to 56 kg N ha) for all treatments. Although CCs increased N2O emissions 21 during the intercrop period compared to F (1.6 and 2.6 times in B and V, respectively), 22 the ISFM resulted in similar cumulative emissions for the CCs and F at the end of the 23 maize cropping period. The higher C:N ratio of the B residue led to a greater proportion 24 1 Biogeosciences Discuss., doi:10.5194/bg-2016-29, 2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c © Author(s) 2016. CC-BY 3.0 License.


Introduction
Improved resource-use efficiencies are pivotal components of a sustainable agriculture that meets human needs and protects natural resources (Spiertz, 2010).
Several strategies have been proposed to improve the efficiency of intensive irrigated systems, where nitrate (NO 3 -) leaching losses are of major concern, both during cash crop and winter fallow periods (Quemada et al., 2013).In this sense, replacing winter intercrop fallow with cover crops (CCs) has been reported to decrease NO 3 -leaching via retention of post-harvest surplus inorganic nitrogen (N) (Wagner-Riddle and Thurtell, 1998), consequently improving N use efficiency (NUE) of the cropping system (Gabriel and Quemada, 2011).Furthermore, the use of CCs as green manure for the subsequent cash crop may further increase soil fertility and NUE (Tonitto et al., 2006;Veenstra et al., 2007) through slow release of N and other nutrients from the crop residues, leading to synthetic fertilizer saving.
From an environmental point of view, N fertilization is closely related with the production and emission of nitrous oxide (N 2 O) (Davidson and Kanter, 2014), a greenhouse gas (GHG) with a molecular global warming potential c. 300 times that of carbon dioxide (CO 2 ) (IPCC, 2007).Nitrous oxide released from agricultural soils is In a CC-maize rotation system, mineral fertilizer application to the cash crop could have an important effect on NUE and N losses from the agro-ecosystem.Different methods for calculating the N application rate (e.g.conventional or integrated) can be employed by farmers, affecting the amount of synthetic N applied to soil and the overall effect of CCs on N 2 O fluxes.Integrated Soil Fertility Management (ISFM) (Kimani et al., 2003) provides an opportunity to optimize the use of available resources, thereby reducing pollution and costs from over-use of N fertilizers (conventional management).
ISFM involves the use of inorganic fertilizers and organic inputs, such as green manure, aiming to maximize agronomic efficiency (Vanlauwe et al., 2011).When applying this technique to a CC-maize crop rotation, N fertilization rate for maize is calculated taking into account the background soil mineral N and the expected available N from mineralization of CC residues, which depends on residue composition.Differences in soil mineral N during the cash crop phase may be significantly reduced if ISFM practices are employed, affecting the GHG balance of the CC-cash crop cropping system.
Only one study has investigated the effect of CCs on N 2 O emissions in Mediterranean cropping systems (Sanz-Cobena et al., 2014).These authors found an effect of CCs species on N 2 O emissions during the intercrop period.After 4 years of CC (vetch, barley or rape)-maize rotation, vetch was the only CC species that significantly enhanced N 2 O losses compared to fallow, mainly due to its capacity to fix atmospheric N 2 and because of higher N surplus from the previous cropping phases in these plots.In this study a conventional fertilization (same N synthetic rate for all treatments) was applied during the maize phase; how ISFM practices may affect these findings remains unknown.Moreover, the relative contribution of mineral N fertilizer, CC residues and/or soil mineral N to N 2 O losses during the cash crop has not been assessed yet.In The study was conducted at "La Chimenea" field station (40°03′N, 03°31′W, 550 m a.s.l.), located in the central Tajo river basin near Aranjuez (Madrid, Spain), where an experiment involving cover cropping systems and conservation tillage has been carried out since 2006.Soil at the field site is a silty clay loam (Typic Calcixerept;Soil Survey Staff, 2014).Some of the physico-chemical properties of the top 0-10 cm soil layer, as measured by conventional methods, were: pH H2O , 8.16; total organic C, 19.0 g kg − 1 ; CaCO 3 , 198 g kg − 1 ; clay, 25%; silt, 49% and sand, 26%.Bulk density of the topsoil layer determined in intact core samples (Grossman and Reinsch, 2002) was 1.46 g cm − 3 .Average ammonium (NH 4 + ) content at the beginning of the experiment was 0.42±0.2mg N kg soil -1 (without differences between treatments).Nitrate concentrations were 1.5±0.2mg N kg soil -1 in fallow and barley and 0.9±0.1 mg N kg soil -1 in vetch.Initial dissolved organic C (DOC) contents were 56.0±7 mg C kg soil -1 in vetch and fallow and 68.8±5 mg C kg soil -1 in barley.The area has a Mediterranean semiarid climate, with a mean annual air temperature of 14 °C.The coldest month is January with a mean temperature of 6 °C, and the hottest month is August with a mean temperature of 24 °C.During the last 30 years, the mean annual precipitation has been approximately 350 mm (17 mm from July to August and 131 mm from September to November).
Hourly rainfall and air temperature data were obtained from a meteorological station located at the field site (CR10X, Campbell Scientific Ltd, Shepshed, UK).A temperature probe inserted 10 cm into the soil was used to measure soil temperature.
Mean hourly temperature data were stored on a data logger.harvesting took place on September 25 th 2014.The fertilizer treatments consisted of AN applied on 2 nd June at three rates: 170, 140 and 190 kg N ha -1 in F, V and B plots, respectively, according to ISFM practices.For the calculation of each N rate, the N available in the soil (which was calculated following soil analysis as described below), the expected N uptake by maize crop, and the estimated N mineralized from V and B residues were taken into account, assuming that crop requirements were 236.3 kg N ha -1 (Quemada et al., 2014) apparent priming effects or different mixing ratios between the added and resident soil N pools) the same amount of N was applied for all treatments.In each subplot, the CC residue was also left on top of the soil.This application took place on 26 th May by spreading the fertilizer homogenously with a hand sprayer, followed by an irrigation event.

Experimental design and agronomic management
Sprinkler irrigation was applied to the maize crop in a total amount of 688.5 mm in 31 irrigation events.Sprinklers were installed in a 12m x 12m framework.The water doses to be applied were estimated from the crop evapotranspiration (ETc) of the previous week (net water requirements).This was calculated daily as ETc.= Kc × ETo, where ETo is reference evapotranspiration calculated by the FAO Penman-Monteith method (Allen et al., 1998) using data from the meteorological station located in the experimental field.The crop coefficient (Kc) was obtained using the relationship for maize in semiarid conditions (Martínez-Cob, 2008).
Two different periods were considered for data reporting and analysis: Period I (from CC sowing to N fertilization of the maize crop), and Period II (from N fertilization of maize to the end of the experimental period, after maize harvest).

GHG emissions sampling and analyzing
al. (2013).One chamber (diameter 35.6 cm, height 19.3 cm) was located in each experimental plot.The chambers were hermetically closed (for 1 h) by fitting them into stainless steel rings, which were inserted at the beginning of the study into the soil to a depth of 5 cm to minimize the lateral diffusion of gases and to avoid the soil disturbance associated with the insertion of the chambers in the soil.The rings were only removed during management events.Each chamber had a rubber sealing tape to guarantee an airtight seal between the chamber and the ring.A rubber stopper with a 3-way stopcock was placed in the wall of each chamber to take gas samples.Greenhouse gas measurements were always made with barley/vetch plants inside the chamber.During the maize period, gas chambers were set up between maize rows.
During Period I, GHGs were sampled weekly or every two weeks.During the first month after maize fertilization, gas samples were taken twice per week.
Afterwards, gas sampling was performed weekly or fortnightly, until the end of the cropping period.To minimize any effects of diurnal variation in emissions, samples were always taken at the same time of the day (10-12 am), that is reported as a representative time (Reeves et al., 2015).
Measurements of N 2 O, CO 2 and CH 4 emissions were made at 0, 30 and 60 min to test the linearity of gas accumulation in each chamber.Gas samples (100 mL) were removed from the headspace of each chamber by syringe and transferred to 20 mL gas vials sealed with a gas-tight neoprene septum.The vials were previously flushed in the field using 80 mL of the gas sample.Samples were analyzed by gas chromatography using a HP-6890 gas chromatograph equipped with a headspace autoanalyzer (HT3), The increases in N 2 O, CH 4 and CO 2 concentrations within the chamber headspace were generally (80% of cases) linear (R 2 > 0.90) during the sampling period (1h).Therefore, emission rates of fluxes were estimated as the slope of the linear regression between concentration and time (after corrections for temperature) and from the ratio between chamber volume and soil surface area (MacKenzie et al., 1998).
Cumulative N 2 O, CH 4 and CO 2 , emissions per plot during the sampling period were estimated by linear interpolations between sampling dates, multiplying the mean flux of two successive determinations by the length of the period between sampling and adding that amount to the previous cumulative total (Sanz-Cobena et al., 2014).The measurement of CO 2 emissions from soil including plants in opaque chambers only includes ecosystem respiration but not photosynthesis (Meijide et al., 2010).Gas samples from the subplots receiving double-labelled AN fertilizer were taken after 60 min static chamber closure 1, 4, 9, 11, 15, 18, 22 and 25 days after fertilizer application.Stable 15 N isotope analysis of N 2 O contained in the gas samples was carried out on a trace gas analyzer (using cryo-trapping and cryo-focusing) coupled to a 20/22 isotope ratio mass spectrometer (both from SerCon Ltd., Crewe, UK), at

15 N Isotope analysis
Rothamsted Research North Wyke.Solutions of 6.6 and 2.9 atom% ammonium sulphate [(NH 4 ) 2 SO 4 ] were prepared and used to generate 6.6 and 2.9 atom% N 2 O (Laughlin et al., 1997) which were used as reference and quality control standards.
During the experiment, the mean natural abundance of atmospheric N 2 O (0.369 atom% 15 N) was subtracted from measured enriched samples to calculate the atom percent excess.To obtain the N 2 O flux that was derived from fertilizer (N 2 O −N dff ), the Eq. ( 1) was used (Loick et al., 2016):

Soil and crop analyses
In order to relate gas emissions to soil properties, soil samples were collected at 0-10 cm depth during the growing season on almost all gas-sampling occasions, particularly after each fertilization event.Three soil cores (2.5 cm diameter and 15 cm Water-Filled Pore Space (WFPS) was calculated by dividing the volumetric water content by total soil porosity.Total soil porosity was calculated according to the relationship: soil porosity = (1-soil bulk density/2.65),assuming a particle density of 2.65 g cm -3 (Danielson and Sutherland, 1986).Gravimetric water content was determined by oven-drying soil samples at 105 °C with a MA30 Sartorius ®.
Four 0.5m × 0.5m squares were randomly harvested from each plot, before killing the CC by applying glyphosate.Aerial biomass was cut by hand at soil level, dried, weighed and ground.A subsample was taken for determination of total N content.
From these samples was determined CC biomass and N contribution to the subsequent maize.
At maize harvest, two 8 m central rows in each plot were collected and weighed in the field following separation of grain and straw.For aboveground N uptake calculations, N content was determined in subsamples of grain and biomass.Total N content on maize and CC subsamples were determined with an elemental analyzer (TruMac CN Leco).Period II (n=11) and the whole experimental period (n=27).Mean soil temperature during the intercrop period was 8.8°C, ranging from 1.8

Cover crop (Period
(December) to 15.5°C (April) (Fig. 1a), which were typical values in the experimental area.Mean soil temperature during maize cropping period was 24.6°C, which was also a standard value for this region.The accumulated rainfall during this period was 215 mm, whereas the 30-year mean is 253 mm.Water-Filled Pore Space ranged from 40 to 81% (Fig. 1b).No significant differences were observed for WFPS mean values between the different treatments (P>0.05).

Mineral N and DOC and cover crop residues
Topsoil NH 4 + content was below 5 mg N kg soil -1 almost of the time in Period I, although a peak was observed after maize sowing (55 days after CCs kill date) (Fig. 2a), with the highest values reached in B (50 mg N kg soil -1 ).Mean NH 4 + content was significantly higher in B than in F (P<0.05).Nitrate content increased after CCs killing, reaching values above 25 mg N kg soil -1 in V treatment (Fig. 2c).Mean NO 3 -content during Period I was significantly higher in the V plots than in the B and F plots (P<0.001).Dissolved Organic C ranged from 60 to 130 mg C kg soil -1 (Fig. 2e).
Average topsoil DOC content was significantly higher in B than in V and F (P<0.05).
The total amount of cover crop biomass left on the ground was 540.5±26.5 and 1106.7±93.6 kg DM ha -1 in B and V, respectively.Accordingly, the total N content of these residues was 11.0±0.6 and 41.3±4.5 kg N ha -1 in B and V, respectively.Nitrous oxide fluxes ranged from -0.06 to 0.22 mg N m -2 d -1 (Fig. 3a) in Period I.The soil acted as a sink for N 2 O at some sampling dates, especially for the F plots.

GHG fluxes
Cumulative fluxes at the end of Period I were significantly greater in CC treatments compared to F (1.6 and 2.6 higher in B and V, respectively) (P<0.05;Table 1).Net CH 4 uptake was observed in all intercrop treatments, and daily fluxes ranged from -0.60 to 0.25 mg C m -2 d -1 (data not shown).No significant differences were observed between treatments in cumulative CH 4 fluxes at the end of Period I (P>0.05;Table 1).Carbon dioxide fluxes (data not shown) remained below 1 g C m -2 d -1 during the intercrop period.Greatest fluxes were observed in B although differences in cumulative fluxes were not significant (P>0.05;Table 1).Nitrous oxide emissions were significantly correlated to CO 2 fluxes (P<0.01,n=17, r=0.69) and soil temperature (P<0.05,n=17, r=0.55).

Environmental conditions and WFPS
Mean soil temperature ranged from 19.6 (reached in September) to 32.3°C (reached in August) with a mean value of 27.9°C (Fig. 1a).Total rainfall during the maize crop period was 57 mm.Water-Filled Pore Space ranged from 19 to 84% (Fig. 1c).Higher mean WFPS values (P<0.01) were measured in B during some sampling dates.Topsoil NH 4 + content increased rapidly after N fertilization (Fig. 2b) decreasing to values below 10 mg N kg soil -1 from 15 days after fertilization to the end of the experimental period.Nitrate concentrations (Fig. 2d) also peaked after AN addition, reaching the highest value (170 mg N kg soil -1 ) 15 days after fertilization in B (P<0.05).

Mineral
No significant differences (P>0.05) between treatments were observed in average soil NH 4 + or NO 3 -during maize phase.Dissolved Organic C ranged from 56 to 138 mg C kg soil -1 (Fig. 2f).Average topsoil DOC content was 26 and 44% higher in B than in V and F, respectively (P<0.001).

GHG fluxes, Yield-Scaled N 2 O emissions and N surplus
Nitrous oxide fluxes ranged from 0.0 to 5.6 mg N m -2 d -1 (Fig. 3b).The highest N 2 O emission peak was observed 1-4 days after fertilization for all plots.Other peaks were subsequently observed until 25 days after fertilization, particularly in B plots where N 2 O emissions 23 and 25 days after fertilization were higher (P<0.05)than those of F and V (Fig. 3b).No significant differences in cumulative N 2 O fluxes were observed between treatments throughout or at the end of the maize crop period (Table 1), albeit fluxes were numerically higher in B than in V (0.05<P<0.10).Daily N 2 O emissions were significantly correlated with NH 4 + topsoil content (P<0.05,n=12, r=0.84).
As in the previous period, all treatments were CH 4 sinks, without significant differences between treatments (P>0.05;Table 1).Respiration rates ranged from 0.15 to 3.0 g C m -2 d -1 ; no significant differences (P>0.05;

Fertilizer-derived N 2 O emissions
The proportion (%) of N 2 O losses from AN, calculated by isotopic analyses, is represented in Fig. 4. The highest percentages of N 2 O fluxes derived from the synthetic fertilizer were observed one day after fertilization, ranging from 34% (V) to 67% (B).
On average, almost 50% of N 2 O emissions in the first sampling event after N synthetic fertilization came from other sources (i.e.soil endogenous N, including N mineralized from the CCs).The mean percentage of N 2 O losses from synthetic fertilizer throughout all sampling dates was 2.5 times higher in B compared to V (P<0.05).There were no significant differences between V and F (P>0.05).

Role of CCs in N 2 O emissions: Period I
Cover crop treatments (V and B) increased N 2 O losses compared to F, especially in the case of V (Table 1).These results are consistent with the meta-analysis of Basche found that V was the only CC significantly affecting N 2 O emissions.The greatest differences between treatments were observed at the beginning (13-40 days after CCs sowing), and at the end of this period (229 days after CCs sowing) (Fig. 3a).On these dates, the mild soil temperatures and the relatively high moisture content were more suitable for soil biochemical processes, which may trigger N 2 O emissions (Fig. 1a, b) (Firestone and Davidson, 1989).Average topsoil NO 3 -was significantly higher in V (Fig. 2b), which was the treatment that led to the highest N 2 O emissions.Legumes such as V are capable of biologically fixing atmospheric N 2 , thereby increasing soil NO 3 content with potential to be denitrified.Further, the mineralization of the most recalcitrant fraction of the previous V residue (which supplies nearly four times more N than the B residue, as indicated in section 3.1.2)together with high C-content sunflower residue could also explain higher NO 3 -contents in V plots (Frimpong et al., 2011), and higher N 2 O losses from denitrification (Baggs et al., 2000).After CCs kill date, N release from decomposition of roots and nodules and faster mineralization of V residue compared to that of B (shown by NO 3 -in soil in Fig. 2c) are the most plausible explanation for the N 2 O increases at the end of the intercrop period (Fig. 3a) (Rochette and Janzen, 2005;Wichern et al., 2008).
Some studies (e.g.Justes et al., 1999;Nemecek et al., 2008) have pointed out that N 2 O losses can be reduced with the use of CCs, due to the extraction of plantavailable N unused by previous cash crop.However, in our study lower N 2 O emissions were measured from F plots without CCs during the intercrop period.This may be a consequence of higher NO 3 -leaching in F plots (Gabriel et al., 2012;Quemada et al., 2013), limiting the availability of the substrate for denitrification.Frequent rainfall during the intercrop period (Fig. 1a) and the absence of N uptake by CCs may have led Nitrous oxide emissions were low during this period, but in the range of those reported by Sanz-Cobena et al. (2014) in the same experimental area.Total emissions during Period I represented 8, 10 and 21% of total cumulative emissions in F, B and V, respectively (Table 1).The absence of N fertilizer application to the soil combined with the low soil temperatures during winterwhich were far from the optimum values for nitrification and denitrification (25-30 °C) processes (Ussiri and Lal, 2012) -may have caused these low N 2 O fluxes.The significant positive correlation between soil temperature and N 2 O fluxes during this period highlights the key role of this parameter as a driver of soil emissions (Schindlbacher et al., 2004;García-Marco et al., 2014).

Role of CCs in N 2 O emissions: Period II
Isotopic analysis during Period II, in which ISFM was carried out, showed that a significant proportion of N 2 O emissions came from endogenous soil N or the mineralization of crop residues, especially after the first days following N fertilization (Fig. 4).In this sense, even though an interaction between crop residue and N fertilizer application has been previously described (e.g. in Abalos et al., 2013), the similar proportion of N 2 O losses coming from fertilizer in B and F (without residue) one day after N fertilization revealed the importance of mineral N harbored in soil micropores in the N 2 O bursts after the first irrigation events.
As we hypothesized, although ISFM practices were adopted, the different CCs played a key role in the N 2 O emissions during Period II.Barley plots had higher N 2 O emissions than fallow or V-residue plots (at the 10% significance level; Table 1).Further, a higher proportion of N 2 O emissions was derived from the fertilizer in Bresidue than in V-residue plots (Fig. 4).These results are in agreement with those of Baggs et al. (2003), who reported a higher percentage of N 2 O derived from the 15 Nlabeled fertilizer using a cereal (ryegrass) as surface mulching instead of a legume (bean).The differences between B and V in terms of cumulative N 2 O emissions and in the relative contribution of each source to these emissions (fertilizer-or soil-N) could be explained by: i) the higher C:N residue of B (20.7±0.7 while that of V was 11.1±0.1,according to Alonso-Ayuso et al. ( 2014)) may have provided an energy source for denitrification (Sarkodie-Addo et al., 2003), increasing the reduction of the NO 3 supplied by the synthetic fertilizer and enhancing N 2 O emissions; ii) NO 3 concentrations, which tended to be higher in B during the maize cropping phase, could have led to incomplete denitrification and larger N 2 O/N 2 ratios (Yamulki and Jarvis, 2002); iii) the easily mineralizable V residue (with low C:N ratio) provided an additional N source for soil microorganisms, thus decreasing the relative amount of N 2 O derived from the synthetic fertilizer (Baggs et al., 2000;Shan and Yan, 2013); and iv) V plots were fertilized with a lower amount of immediately available N (i.e.AN) than B plots, which could have resulted in better synchronization between N release and crop needs (Ussiri and Lal, 2012) in V plots.Supporting these findings, Bayer et al. (2015) recently concluded that partially supplying the maize N requirements with winter legume cover-crops can be considered a N 2 O mitigation strategy in subtropical agroecosystems.
The mineralization of B residues resulted in higher DOC contents for these plots compared to the F or V plots (P<0.001).This was observed in both Period I (as a consequence of soil C changes after the 8-year cover-cropping management microbial N immobilization (Frimpong andBaggs, 2010, Dendooven et al., 2012).In our experiment, a N 2 O peak was observed in B plots 20-25 days after fertilization (Fig. 3b) after a remarkable increase of NO 3 -content (Fig. 2d), which may be a result of a remineralization of previously immobilized N in these plots.
The positive correlation of N 2 O fluxes and soil NO 3 -content and WFPS during the whole cycle further supports the importance of denitrification process for explaining N 2 O losses in this agro-ecosystem (Davidson et al., 1991;García-Marco et al., 2014).
However, the strong positive correlation of N 2 O with NH 4 + indicated that nitrification was also a major process leading to N 2 O fluxes, and showed that the continuous dryingwetting cycles during a summer irrigated maize crop in a semi-arid region can lead to favorable WFPS conditions for both nitrification and denitrification processes (Fig. 1c) (Bateman and Baggs, 2005).Emission Factors ranged from 0.2 to 0.6% of the synthetic N applied, which were lower than the IPCC default value of 1%.As explained above, ecological conditions during the intercrop period (rainfall and temperature) and maize phase (temperature) could be considered as normal (based on the the 30-year average) in Mediterranean areas.Aguilera et al. (2013) obtained a higher emission factor for high (1.01%) and low (0.66%) water-irrigation conditions in a meta-analysis of Mediterranean cropping systems.As is generally found in non-flooded arable soils, all treatments were net CH 4 sinks (Snyder et al., 2009).No significant differences were observed between treatments in any of the two periods (Table 1), which is similar to the pattern observed by Sanz-Cobena et al. (2014).Some authors (Dunfield and Knowles, 1995;Tate, 2015) have suggested an inhibitory effect of soil NH 4 + on CH 4 uptake.Low NH 4 + contents during almost all of the CCs and maize cycle may explain the apparent lack of this inhibitory effect (Banger et al., 2012).However, during the dates when the highest NH 4 + contents were reached in V and B (225 days after CCs sowing) (Fig. 3a), CH 4 emissions were significantly higher for these plots (0.12 and 0.16 mg CH 4 -C m -2 d -1 for V and B, respectively) than for F (-0.01 mg CH 4 -C m -2 d -1 ) (data not shown).Similarly, the NH 4 + peak observed two days after fertilization (Fig. 3b) decreased in the order V>F>B, the same trend as CH 4 emissions (which were 0.03, -0.04 and -0.63 mg CH 4 -C m -2 d -1 in V, F and B, respectively; data not shown).Contrary to Sanz-Cobena et al. (2014), the presence of CCs did not increase CO 2 fluxes (Table 1) during Period I (which was longer than that considered by these authors), even though higher fluxes tended to be associated to B plots, probably as a consequence of higher root biomass and plant respiration rates in the cereal (B) than in the legume (V).The decomposition of CC residues and the growth of maize rooting system resulted in an increase of CO 2 fluxes during Period II (Oorts et al., 2007;Chirinda et al., 2010), although differences between treatments were not observed.

Yield-scaled emissions, N surplus and general assessment
Yield-scaled N 2 O emissions ranged from 1.74 to 7.15 g N 2 O-N kg aboveground N uptake -1 , which is about 1-4 times lower than those reported in the meta-analysis of - order to confirm the potential of CCs for increasing both the agronomic and environmental efficiency of irrigated cropping areas.

Conclusions
Our study confirmed that the presence of CCs (particularly V) during the intercrop period increased N 2 O losses, but the contribution of this phase to cumulative        Vertical lines indicate standard errors."NS" and * denote not significant and significant at P<0.05, respectively.
Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License.Twelve plots (12m × 12m) were randomly distributed in four replications of three cover cropping treatments, including a cereal and a legume: 1) barley (B) (Hordeum vulgare L., cv.Vanessa), 2) vetch (V) (Vicia sativa L., cv.Vereda), and 3) traditional winter fallow (F).Cover crop seeds were broadcast by hand over the stubble of the previous crop and covered with a shallow cultivator (5 cm depth) on October 10 th 2013, at a rate of 180 and 150 kg ha − 1 for B and V, respectively.The cover cropping phase finished on March 14 th 2014, with an application of glyphosate (Nphosphonomethyl glycine) at a rate of 0.7 kg a.e.ha -1 .All the CC residues were left on top of the soil.Thereafter, a new set of N fertilizer treatments was set up for the maize cash crop phase.Maize (Zea mays L., Pioneer P1574, FAO Class 700) was direct drilled on April 7 th 2014 in all plots, resulting in a plant population density of 7.5 plants m −2 ; both from Agilent Technologies (Barcelona, Spain).HP Plot-Q capillary columns Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License.transported gas samples to a 63 Ni electron-capture detector (Micro-ECD) to analyze N 2 O concentrations and to a flame ionization detector (FID) connected to a methanizer to measure CH 4 and CO 2 (previously reduced to CH 4 ).The temperatures of the injector, oven and detector were 50, 50 and 350ºC, respectively.The accuracy of the gas chromatographic data was 1% or better.Two gas standards comprising a mixture of gases (high standard with 1500 ± 7.50 ppm CO 2 , 10 ± 0.25 ppm CH 4 and 2 ± 0.05 ppm N 2 O and low standard with 200 ± 1.00 ppm CO 2 , 2 ± 0.10 ppm CH 4 and 200 ± 6.00 ppb N 2 O) were provided by Carburos Metálicos S.A. and Air Products SA/NV, respectively, and used to determine a standard curve for each gas.The response of the GC was linear within 200-1500 ppm for CO 2 and 2-10 ppm CH 4 and quadratic within 200-2000 ppb for N 2 O.
Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License.length) were randomly sampled close to the ring in each plot, and then mixed and homogenized in the laboratory.Soil NH 4 + and NO 3 -concentrations were analyzed using 8 g of soil extracted with 50 mL of KCl (1 M), and measured by automated colorimetric determination using a flow injection analyzer (FIAS 400 Perkin Elmer) provided with a UV-V spectrophotometer detector.Soil (DOC) was determined by extracting 8 g of homogeneously mixed soil with 50 mL of deionized water, and analyzed with a total organic C analyser (multi N/C 3100 Analityk Jena) equipped with an IR detector.The Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License.Yield-scaled N 2 O emissions and N surplus in the maize cash crop were calculated as the amount of N 2 O emitted (considering the emissions of the whole experiment, i.e.Period I + Period II) per unit of above-ground N uptake, and taking the difference between N application and above-ground N uptake, respectively (van Groenigen et al., 2010).Statistical analyses were carried out with Statgraphics Plus 5.1.Analyses of variance were performed for all variables over the experiment (except climatic ones), for both periods indicated in section 2.2.Data distribution normality and variance uniformity were previously assessed by Shapiro-Wilk test and Levene's statistic, respectively, and transformed (log10, root-square, arcsin or inverse) before analysis when necessary.Means of soil parameters were separated by Tukey's honest significance test at P<0.05, while cumulative GHG emissions, YSNE and N surplus were compared by the orthogonal contrasts method at P<0.05.For non-normally distributed data, the Kruskal-Wallis test was used on non-transformed data to evaluate differences at P<0.05.Linear correlations were carried out to determine relationships between gas fluxes and WFPS, soil temperature, DOC, NH 4 + and NO 3 -.Theses analyses were performed using the mean/cumulative data of the replicates of the CC treatments (n=12), and also for all the dates when soil and GHG were sampled, for Period I (n=16), I) 3.1.1Environmental conditions and WFPS 13 Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License.
et al. (2014), which showed that overall CCs increase N 2 O fluxes (compared to bare fallow), with highly significant increments in the case of legumes and a lower effect in Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License. the case of non-legume CCs.In the same experimental area, Sanz-Cobena et al. (2014) Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License. to N losses through leaching, resulting in low concentrations of soil mineral N in F plots.
) and Period II (due to the CC decomposition).Although in the present study the correlation between Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License.

N 2 O
emissions considering the whole cropping cycle (intercrop-cash crop) was low (8-21%).The high influence of the maize crop period over total N 2 O losses was not only due to N synthetic fertilization, but also to CC residue mineralization and especially endogenous soil N. The type of CC residue determined the N synthetic rate in a ISFM system and affected the percentage of N 2 O losses coming from N fertilizer/soil N as well as the pattern of N 2 O losses during the maize phase (through changes in soil NH 4 + , NO 3 -and DOC concentrations).By employing ISFM, similar N 2 O emissions were measured from CCs and F treatments at the end of the whole cropping period, resulting in low YSNE (3-6 g N 2 O-N kg aboveground N uptake -1 ) and N surplus (31 to 56 kg N ha -1 ).Replacing winter F by CCs did not affect significantly CH 4 uptake or respiration rates neither during intercrop or maize cropping periods.Our results highlight the critical importance of the cash crop period on total N 2 O emissions, and demonstrate that the use of either legume or non-legume CC combined with ISFM may provide an optimum balance between GHG emissions from crop production and agronomic efficiency.
fluxes, yield-scaled N 2 O emissions (YSNE) and N surplus in the three cover crop treatments (fallow, F, 772 vetch, V, and barley, B) at the end of both cropping periods.P value was calculated with Student's t-test and d.f.=9.(*) and S.E.denote significant at P<0.05 773 and the standard error of the mean, respectively.

Figure captions : Figure 1 .
Figure captions: Figure 1.Daily mean soil temperature (°C) rainfall and irrigation (mm) (a) and soil WFPS (%) in the three cover crop (CC) treatments (fallow, F, vetch, V, and barley, B) during Period I (b) and II (c).Vertical lines indicate standard errors.

Figure
Figure 2a, b NH 4 + -N; c, d NO 3 --N; and e, f DOC concentrations in the 0-10 cm soil layer for the three cover crop (CC) treatments (fallow, F, vetch, V, and barley, B) during both cropping periods.The black arrows indicate the time of spraying glyphosate over the cover crops.The dotted arrows indicate the time of maize sowing.Vertical lines indicate standard errors.

Figure 3 .
Figure 3. N 2 O emissions for the three cover crop (CC) treatments (fallow, F, vetch, V, and barley, B) during Period I (a) and II (b).The black arrows indicate the time of spraying glyphosate over the cover crops.The dotted arrows indicate the time of maize sowing.Vertical lines indicate standard errors.

Figure 4 .
Figure 4. Proportion of N 2 O losses (%) coming from N synthetic fertilizer during Period II, for the three cover crop treatments (fallow, F, vetch, V, and barley, B).
van Groenigen et al. (2010)i:10.5194/bg-2016-29,2016ManuscriptunderreviewforjournalBiogeosciencesPublished:29 March 2016 c Author(s) 2016.CC-BY 3.0 License.vanGroenigenetal.(2010) for a fertilizer N application rate of 150-200 kg ha -1 .Mean N surpluses of V and F (Table1) were in the recommended range (0-50 kg N ha −1 ) byvan Groenigen et al. (2010), while the mean N surplus in B (55 kg N ha −1 ) was also close to optimal.In spite of higher N 2 O emissions in V during Period I (which accounted for a low proportion of total cumulative N 2 O losses during the experiment), these plots did not emit greater amounts of N 2 O per kg of N taken up by the maize plants, and even tended to decrease YSNE and N surplus (Table1).
(Sanz-Cobena et al., 2012;Adviento-Borbe et al., 2007)o lower N 2 O fluxes than previous experiments where conventional N rates were applied(Sanz-Cobena et al., 2012;Adviento-Borbe et al., 2007), in agreement with the study of Migliorati et al. 1 with the introduction of CCs.These environmental factors together with CO 2 emissions associated to CCs sowing and killing, should be assessed in future studies in Biogeosciences Discuss., doi:10.5194/bg-2016-29,2016 Manuscript under review for journal Biogeosciences Published: 29 March 2016 c Author(s) 2016.CC-BY 3.0 License.

Table 1
Total cumulative N