**4. Discussion**

This study evaluated the effect on GHG emissions from soil under three different agricultural managemen<sup>t</sup> systems, an integrated (INT), an organic (ORG) and an organic conservation (ORG+) system, on an irrigated vegetable crop rotation for two years, and the relationship of GHG fluxes with soil variables.

Daily fluxes of N2O correlated positively with soil temperature and CO2 fluxes, probably caused by the high microbial activity associated to the organic matter mineralization in the warm season. Indeed, higher peaks in N2O emissions occurred mainly between the end of March and the beginning of October, namely the period with higher soil temperatures (>20 ◦C). Other studies reported that soil temperature may be a driver for N2O production when substrates are abundant, and the soil water content is optimal for microbial processes [32,33]. However, in our experiment the period with higher soil temperatures corresponded to that during which all N fertilization occurred, thus, it is difficult to consider separately the effect of the two drivers on N2O emissions.

The agricultural managemen<sup>t</sup> system influenced the average daily N2O flux within F1 in P1 and P2, and within F2 in P2; in those cases, we found lower values in ORG+ than in the other systems, likely due to the significantly lower N fertilizer rate supplied to ORG+. Indeed, ORG+ had significantly lower cumulative N2O emissions than INT and ORG in both fields in period 2, in which no fertilizers were supplied to spring and summer lettuce in ORG+.

We did not find a significant correlation between nitrate concentration in soil and N2O emissions, even if nitrates were higher after mineral fertilization events in INT than organic fertilization in ORG and ORG+ in few sampling days in summer 2015, since the low number of the soil samples (29 in F1 and 27 in F2) may have negatively affected the robustness of the model.

The effect of nitrogen fertilization events, implying either mineral or organic N forms, on stimulating both short-term N2O flux and cumulative N2O emissions, was already reported by many authors [34,35]. In our study, high peaks of N2O (>10 mg N2O m<sup>−</sup><sup>2</sup> day−1) were recorded a few days after fertilization events (4–10 days), in accordance to what was reported by Volpi et al. [36] in a similar soil and in the same environment.

In our study, peaks on daily N2O flux were generally higher (>15 mg N2O m<sup>−</sup><sup>2</sup> day−1) after organic N fertilization events (ORG and ORG+), than after mineral N fertilization (INT). The occurrence

of peaks in soil N2O emissions after the application of organic fertilizers have been explained by other studies [37,38], as an e ffect of the increased availability of N and C for the soil microbial community. Thus, the increased microbial activity leads to high O2 consumption that may create anaerobic conditions suited for the denitrification process from which N2O is originated.

Di fferently, other studies reported lower N2O emissions with organic fertilizers than mineral fertilizers, especially with solid manure, due to a slower release of N respect to mineral fertilizers or liquid slurry [14]. However, the e ffect of fertilizers on soil GHG emissions strictly depends on climate and soil specific conditions as well as on the type of the organic fertilizer itself. In fact, Pelster et al. [39], comparing four di fferent N sources (one mineral fertilizer and three di fferent manures), observed that N2O emissions responded similarly to organic and mineral N sources in high-C soils, whereas in low-C soils N2O emissions may be specifically stimulated by the use of C-rich manures. Moreover, the application technique of organic fertilizers may influence the soil N2O emissions. Indeed, the incorporation of organic fertilizers is expected to increase N2O emissions when soil moisture status is suitable for N2O production, while ammonia volatilization may decrease, since more N entered the soil [40]. However, in our experiment we highlighted a tendency for lower N2O emissions in ORG+ where the fertilizers were broadcasted more on soil surface than in ORG, where they were incorporated in soil, though that result was most probably due to the low N rate applied in ORG+.

Moreover, peaks of N2O emissions, in a range between 5 and 20 mg N2O m<sup>−</sup><sup>2</sup> day−1, occurred from 10 to 15 days after the soil incorporation of the green manures in ORG. Heller et al. [41] in Mediterranean conditions, recorded the highest N2O flux maximum two weeks after the tillage operations practiced for maize residues incorporation. Other authors reported that the incorporation of crop biomass into the soil produced N2O and CO2 peaks due to the increased availability of substrates for mineralization and microbial activity, when soil moisture was not limiting [42,43]. In particular, it was reported that N2O emissions are generally increased when crop biomass with a low C:N ratio is incorporated in the soil [17,18]. However, in our study, peaks in N2O emissions were similar (15–20 mg N2O m<sup>−</sup><sup>2</sup> day−1) after the incorporation of both spring green manure, composed by only legumes (C:N = 12, Tables S1 and S2) and summer green manure, composed by one legume, two cereals and one pseudo-cereal (C:N = 33, Tables S1 and S2). Indeed, peaks in N2O emissions might have been due to an improvement of C availability in soil after plant material incorporation that stimulated denitrification [44].

Daily fluxes of CH4 were negative in about 80% and 90% of the sampling days in F1 and F2, respectively. CH4 uptake by soil was similar in all the cropping systems, with higher uptakes recorded only in the ORG system in F2 (average −0.19 mg CH4 m<sup>−</sup><sup>2</sup> day−1). Values of CH4 uptake recorded in our experiment were in the range reported by literature for non-flooded agricultural soils (from 0 to 1.03 mg CH4 m<sup>−</sup><sup>2</sup> day−1) [22]. However, CH4 uptake was lower than that reported by Flessa et al. [45] on a potato field in a temperate climate (average −0.35 mg CH4 m<sup>−</sup><sup>2</sup> day−1). Cumulative CH4 emissions were not di fferent among the cropping systems, in both periods and fields. In that regard, our results comply with other studies that reported no e ffect by conservation tillage on CH4 emissions [46]. Di fferently, other studies comparing organic and non-organic managemen<sup>t</sup> revealed a slightly, but significantly higher, net CH4 uptake in organic cropping systems [47]. Moreover, the higher mineral fertilizer rate distributed in INT and the higher tillage intensity of INT and ORG seemed to have not inhibited the soil CH4 oxidation capacity; namely the methanotrophic activity of microorganisms in soil, compared to the ORG+ system. However, the recent implementation of the three cropping systems could not ye<sup>t</sup> have a ffected the soil stability, as well as the gas di ffusion and the methanotrophic activity in soil that may influence CH4 uptake [22]. Indeed, the number of years since the initiation of conservation tillage is a key issue for evaluating and understanding the e ffects generated by this managemen<sup>t</sup> strategy [48].

Moreover, our study showed no di fferences in CH4 emissions during periods of bare fallow (INT) and periods with cover crops (ORG and ORG+), similarly to what was reported by Sanz-Cobena et al. [49] and Guardia et al. [50] in a maize/cover crop rotation. However, studies are scarce on this topic, thus further research is needed to investigate the e ffect of cover crops on CH4 emissions [16].

Concerning soil conditions, we did not find any strong correlations among soil temperature, WFPS and CH4 emissions, with only a weak positive correlation between soil CH4 emissions and WFPS in ORG. In our experiment WFPS did not show prevalently very low or high values, and soil water content was not as a strong driver for CH4 emissions as reported in other studies, where it lowered the activity of methanotrophic bacteria in very dry or very wet soil conditions [51].

Measurements of daily CO2 flux in our experiment ranged from 3.9 to 65.2 g CO2 m<sup>−</sup><sup>2</sup> day−1, with values often higher than in other studies conducted in a Mediterranean environment on fertilized crops, including organic cultivation or cover crops (1.5–25.7 g CO2 m<sup>−</sup><sup>2</sup> day−1) [49,52]. In all treatments, the intensity of CO2 daily fluxes followed the variations of soil temperature, with values generally higher (up to 60 g CO2 m<sup>−</sup><sup>2</sup> day−1) during the warm season, between April and September, than in the rest of the year (<25 g CO2 m<sup>−</sup><sup>2</sup> day−1). Our results confirm the positive relationship between soil temperature and CO2 flux, usually non-linear, reported by other authors [53–55]. In our experiment, irrigation may have contributed to the high values of CO2 daily flux measured during the warm season, compared to those of other studies conducted in drought stressed Mediterranean environments. Indeed, Almagro et al. [56] reported that soil respiration varied following changes in soil moisture in late spring and summer, in a dry meso-Mediterranean climate, and that soil respiration was strongly limited by soil water content (SWC) < 10%. In our study, irrigation allowed us to maintain soil water content above 9% (20% WFPS), with the exception of three dates. In such a condition, soil water was never limiting for biological processes deputed to the production of CO2, including root respiration. Our results highlighted a negative correlation between WFPS and CO2 daily flux, only due to the stronger positive correlation of CO2 daily flux and soil temperature and to the inverse pattern of WFPS and soil temperature values both in winter and in summer periods.

Furthermore, our results showed a di fferent e ffect of the cropping systems on daily flux of CO2, as the intercept of the linear regression describing the relationship between CO2 flux and soil temperature was higher in ORG and ORG+ than in INT. Thus, besides the variation mediated by soil temperature and water content, the level of organic substrates supplied to the soil in ORG and ORG+ have determined higher soil respiration rates.

Moreover, the incorporation of soil of green manure (ORG) might have been a significant driver for short-term CO2 fluxes, due to the proneness of green manure to mineralization [52,57]. Indeed, CO2 daily flux was higher in ORG than in the other treatments a few days after the green manure incorporation was carried out, before summer lettuce cultivation in F1, and before fennel cultivation in F2.

Short-term peaks in CO2 daily flux were also recorded in F1 after main tillage for sowing green manure and cabbage transplanting, and in F2 after tillage for fennel transplanting. Peaks in CO2 emissions after main tillage were previously reported by many authors, mainly due to an increased mineralization of soil organic matter, as well as a transitory e ffect, due to the removal of physical constraints on CO2 di ffusion [46,58]. Cumulative CO2 emissions ranged between 2.0 and 8.2 t C–CO2 ha−<sup>1</sup> and when there was a significant di fference among the cropping systems, we highlighted a tendency in higher emissions in ORG and/or ORG+ than in INT. In ORG, the green manure incorporation and the organic fertilizer application could have increased the soil heterotrophic respiration [41] as discussed above, while in ORG+, living mulch may have increased the autotrophic component of respiration [59,60]. These results are in line with Chirinda et al. [61] that reported an increase of CO2 emissions due both to manure application and to catch crops' cultivation in a sandy loam soil.

The net GHG emissions budget showed a tendency of being higher in ORG+ (in both fields and periods) and ORG (in P1 in F1 and F2) with respect to INT because of the e ffect of the cropping system on CO2 emissions, since the CO2-eq of non-CO2 GHG were not di fferent among INT, ORG and ORG+.

Thus, the integration of organic and conservation agriculture showed a tendency of higher CO2 emissions and lower N2O emissions than the other cropping systems, with no clear potential for soil GHG mitigation, at least in the first two years of organic conservation management. Indeed, a

long-term field trial could help to clarify whether the result of this study on the e ffect of ORG+ on soil CO2 and N2O emissions was only transitory, especially considering the importance of the duration of no-till [62]. It is well known, indeed, that the introduction of no-till practices may require a long time to produce beneficial e ffects on soil's physical and biological aspects, which may bu ffer the GHG emission potential of the soil.

In the transition phase, a possible solution to improve the distribution of fertilisers in the soil profile and to sustain crop yield could come from a di fferent fertilization strategy. The within-furrow application of organic fertilizers at transplant—which could be possible by means of a fertilizer tank mounted on the direct transplanting machine—and fertigation with organic material, may result in a better stratification of fertilizers even in no-till conditions, allowing the reduction of the exposure of organic fertilisers to oxidation conditions, while increasing their e fficiency.

Moreover, the trade-o ff between GHG mitigation and the crop productivity has to be taken into account, evaluating the crop yield in the three cropping systems [63].
