Effect of Organic Amendment Addition on Soil Properties, Greenhouse Gas Emissions and Grape Yield in Semi-Arid Vineyard Agroecosystems

: In semi-arid vineyard agroecosystems, highly vulnerable in the context of climate change, the soil organic matter (OM) content is crucial to the improvement of soil fertility and grape productivity. The impact of OM, from compost and animal manure, on soil properties (e.g., pH, oxidisable organic C, organic N, NH 4+ -N and NO 3 − -N), grape yield and direct greenhouse gas (GHG) emission in vineyards was assessed. For this purpose, two wine grape varieties were chosen and managed differently: with a rain-fed non-trellising vineyard of Monastrell, a drip-irrigated trellising vineyard of Monastrell and a drip-irrigated trellising vineyard of Cabernet Sauvignon. The studied fertiliser treatments were without organic amendments (C), sheep/goat manure (SGM) and distillery organic waste compost (DC). The SGM and DC treatments were applied at a rate of 4600 kg ha − 1 (fresh weight, FW) and 5000 kg ha − 1 FW, respectively. The use of organic amendments improved soil fertility and grape yield, especially in the drip-irrigated trellising vineyards. Increased CO 2 emissions were coincident with higher grape yields and manure application (max-imum CO 2 emissions = 1518 mg C-CO 2 m − 2 d − 1 ). In contrast, N 2 O emissions, mainly produced through nitriﬁcation, were decreased in the plots showing higher grape production (minimum N 2 O emissions = − 0.090 mg N 2 O-N m − 2 d − 1 ). In all plots, the CH 4 ﬂuxes were negative during most of the experiment ( − 1.073 − 0.403 mg CH 4 -C m − 2 d − 1 ), indicating that these ecosystems can represent a signiﬁcant sink for atmospheric CH 4 . According to our results, the optimal vineyard management, considering soil properties, yield and GHG mitigation together, was the use of compost in a drip-irrigated trellising vineyard with the grape variety Monastrell. This study was conducted during the 2014 crop season in three commercial wine grape vineyard plots close to in the The


Introduction
Vineyard soils covered a total area of 7.5 Mha worldwide in 2016. In that year, Spain was the country with the greatest extension of vineyards (975,270 ha) and had a wine production of around 40 × 10 6 hL [1]. Most of the Spanish vineyards are concentrated in the Mediterranean basin, now highly vulnerable in the context of climate change [2], with viticulture being a significant economic sector (the wine production generated an income of €4.8 billion in 2016) [3] and part of the socio-cultural heritage in this zone. Appropriate soil management practices are needed to ensure the environmental sustainability of the viticulture sector, a situation increasingly demanded by consumers [4].
The management of the vineyards includes various agricultural practices affecting soil functions. The most common soil management techniques in Mediterranean vineyards

Study Area and Soil Characteristics
This study was conducted during the 2014 crop season in three commercial wine grape vineyard plots close to Monovar, in the Vinalopo region (Alicante-south-eastern Spain; 38o27 4 N, 0o54 56 W; elevation 392 m a.s.l.). The area has a semi-arid continental Mediterranean climate, with a mean annual temperature of 15.1 • C, mean annual rainfall of 345 mm and an annual Penman-Monteith Evapotranspiration of 1400 mm (average value based on data collected from 2005 to 2015 by [26]). Figure 1 shows the precipitation and air temperatures recorded during the period studied. The soil in this area has a sandy-loam texture and is classified as a Xeric Haplocalcids [27]. The plots studied had soils with similar characteristics, with 8%-14% silt, 9%-15% clay and 71%-81% sand. They were calcareous (118-189 g of active CaCO 3 kg −1 ) and had values of pH and electrical conductivity (EC) in water extracts of 8.3-8.4 and 0.26-0.28 dS m −1 , respectively. The oxidisable organic carbon (oxidisable OC) contents were low (3.3-4.4 g kg −1 ). dynamics, grape yield and direct GHG emissions of Mediterranean wine grape vineyards managed with different practices-in terms of the training system, irrigation, N fertiliser and grape variety-over one full growing season.

Study Area and Soil Characteristics
This study was conducted during the 2014 crop season in three commercial wine grape vineyard plots close to Monovar, in the Vinalopo region (Alicante-south-eastern Spain; 38o27′4′'N, 0o54′56′'W; elevation 392 m a.s.l.). The area has a semi-arid continental Mediterranean climate, with a mean annual temperature of 15.1 °C, mean annual rainfall of 345 mm and an annual Penman-Monteith Evapotranspiration of 1400 mm (average value based on data collected from 2005 to 2015 by [26]). Figure 1 shows the precipitation and air temperatures recorded during the period studied. The soil in this area has a sandyloam texture and is classified as a Xeric Haplocalcids [27]. The plots studied had soils with similar characteristics, with 8%-14% silt, 9%-15% clay and 71%-81% sand. They were calcareous (118-189 g of active CaCO3 kg −1 ) and had values of pH and electrical conductivity (EC) in water extracts of 8.3-8.4 and 0.26-0.28 dS m −1 , respectively. The oxidisable organic carbon (oxidisable OC) contents were low (3.3-4.4 g kg −1 ). The plots had two different varieties of grape (Vitis vinifera L.), namely 15-year-old Monastrell and 10-year-old Cabernet Sauvignon, grafted onto 110 Richter rootstock. The training systems were a bilateral cordon trellised to a three-wire vertical system (dripirrigated) and the traditional "bush vine" method, without a system of support (rain-fed vineyard). The plantation densities were 2000 and 2300 vines per hectare for non-trellised and trellised grapevines, respectively.

Experimental Design
This work involved the study of three treatments, with three replicate sub-plots per treatment (each containing 10 planted vines) set up in a completely randomised design. The treatments were two organic amendments, sheep/goat manure (SGM) and distillery organic waste compost (DC), along with an unfertilised control (C). Table 1 shows the physico-chemical properties and chemical composition of the organic materials used. The application of the amendments was adjusted to provide 170 kg ha −1 of N, the rate allowed for organic N amendments in zones vulnerable to groundwater pollution by nitrate from agricultural sources [28]. Based on this, the application rate was 4600 kg ha −1 FW of SGM and 5000 kg ha −1 FW of DC. These amendments were applied to the surface and incorporated into the soil to a depth of 30 cm in March, using a chisel plough. Synthetic fertiliser The plots had two different varieties of grape (Vitis vinifera L.), namely 15-year-old Monastrell and 10-year-old Cabernet Sauvignon, grafted onto 110 Richter rootstock. The training systems were a bilateral cordon trellised to a three-wire vertical system (dripirrigated) and the traditional "bush vine" method, without a system of support (rain-fed vineyard). The plantation densities were 2000 and 2300 vines per hectare for non-trellised and trellised grapevines, respectively.

Experimental Design
This work involved the study of three treatments, with three replicate sub-plots per treatment (each containing 10 planted vines) set up in a completely randomised design. The treatments were two organic amendments, sheep/goat manure (SGM) and distillery organic waste compost (DC), along with an unfertilised control (C). Table 1 shows the physico-chemical properties and chemical composition of the organic materials used. The application of the amendments was adjusted to provide 170 kg ha −1 of N, the rate allowed for organic N amendments in zones vulnerable to groundwater pollution by nitrate from agricultural sources [28]. Based on this, the application rate was 4600 kg ha −1 FW of SGM and 5000 kg ha −1 FW of DC. These amendments were applied to the surface and incorporated into the soil to a depth of 30 cm in March, using a chisel plough. Synthetic fertiliser was not applied during the experiment and only three supplemental drip irrigations (100 mm each) were carried out-on 15 May, 21 July and 25 August 2014-in the plots with Agronomy 2021, 11, 1477 4 of 20 trellised grapevines. All plots had reduced tillage and the usual pesticide treatments were applied (Figures 1 and 2).

Soil Analysis and Grape Yield Determination
In all sub-plots of each plot, topsoil samples were collected (0-30 cm depth) to evaluate the effects of the studied treatments on soil physico-chemical and chemical proper-

Soil Analysis and Grape Yield Determination
In all sub-plots of each plot, topsoil samples were collected (0-30 cm depth) to evaluate the effects of the studied treatments on soil physico-chemical and chemical properties. In all replicates of each treatment, six sub-samples were taken along the row of 10 vines, three on each side; then, these sub-samples were combined to compose a single sample. Soil samples were collected 1, 85, 190 and 296 days after the amendment application (24 March,17 June and 29 September 2014 and 13 January 2015). Each soil sample was air-dried and sieved to 2 mm, after removal of vegetation and stones.
The pH, EC, texture, active CaCO 3 , oxidisable OC and distinct forms of N (total Kjeldahl, nitrate and ammonium N) were determined in soil samples according to the methods used by Medina et al. [29]. The characteristics and composition of the organic materials used in this experiment were determined according to Idrovo-Novillo et al. [30]. All analyses of the organic amendment and soil samples were made in triplicate.
On the days of harvest, 10 and 15 October 2014, all grape clusters of each plant contained in each sub-plot (10 vines) were weighed to determine the yield per vine, on a fresh weight basis.

GHG Determinations
Opaque static chambers with a volume of 18.3 L (diameter 36 cm, height 18 cm), and with a design similar to that described in Recio et al. [31], were used to sample the GHG fluxes. There were two chambers in each sub-plot (10 vines), adjusted to a metallic frame placed in the soil prior to the gas measurements to avoid the effect of soil disturbances on GHG fluxes. A rubber stopper with a 3-way stopcock was placed in the top of each chamber so that gas samples could be taken. Gas samples were taken 0, 20, 40 and 60 min after chamber closure, by the gas pooling technique [32]. An additional syringe was connected and was pumped several times before sampling the gas fluxes, to achieve homogenous mixing of the headspace air. A 30-mL gas sample was collected from each of the two chambers in each sub-plot, with the same syringe, at each sampling time. Each combined 60-mL sample was then immediately transferred to a 20-mL gas chromatography vial sealed with a gas-tight neoprene septum; 40 mL were injected with recirculation to flush the vial and overpressure was reached with the remaining 20 mL. To minimise the effects of diurnal variation on the composition of the samples, they were taken at the same time of day (9 a.m. to 12 a.m.) in each sampling.
The change in gas concentration within the chamber during the 60 min of closure was used to calculate the flux rates of N 2 O, CH 4 and CO 2 . The linearity of gas diffusion into the chamber headspace during this closure period was determined previously, and each flux was calculated from a single determination at the end of the closure, transforming the gas measurement from N 2 O, CH 4 and CO 2 ppm (µL/L) to mg N 2 O-N, CH 4 -C and CO 2 -C m −2 d −1 , respectively. For this calculation, the volume of the chamber (18.3 L) and the surface area of the soil (0.108 m 2 ) were considered [34]. Measurements of GHGs (CO 2 , N 2 O and CH 4 ) were made during the experimental period (days 1, 7, 15, 42, 85, 106, 190, 242 and 296 after the amendment application, corresponding to 24 and 31 March, 8 April, 5 May, 17 June, 7 July, 29 September and 20 November 2014 and 13 January 2015). The cumulative gas emissions during the experimental period were calculated by multiplying the average flux of two successive determinations by the length of time between them; this amount was added to the previous cumulative total [35].

Statistical Methods
SPSS v. 22.0 statistical software was used to carry out the statistical analysis. The standard deviation was determined for the mean values of each parameter measured in the soil. Statistical comparisons of the treatment means for the grape production and cumulative gas emissions were performed using one-way ANOVA and the separation of these means was established by the Tukey-b test. Finally, a factorial analysis (FA) was used to determine the correlations among the mean values of all the parameters measured in the soil. An FA shows the correlated variables as a new set of uncorrelated, mutually orthogonal variables, each of which is a linear combination of the original variables. The new, calculated variables are called 'factorial components' (FCs). The FA was performed using the mean values of the three plots for each treatment. These mean values were calculated from the data of the different days on which the soil parameters and GHG emissions were determined (24 March, 17 June and 29 September 2014 and 13 January 2015). The factorial loads of the data were analysed after the application of the Varimax normalised rotation to the coordinate system of the FCs.

Effect of the Organic Fertiliser Type on Soil Parameters and Grape Yield
In most cases, soil pH showed an increase throughout the experiment ( Figure 3A). However, this increase was, in general, lower in the amended soils than in the control soil; a lower soil pH could reduce vine chlorosis (common in this zone) by improving the assimilation of plant micronutrients, such as Fe. Similarly, Hannachi et al. [36] found that the addition of organic amendments produced a decrease in the pH of alkaline cultivated soils in semi-arid areas of southern Tunisia, considering this result as beneficial with regard to increasing the availability of nutrients to plants. These authors indicated that the oxidation of NH 4 + from N fertilisers (to produce nitrate), the higher root uptake of cations with respect to anions and the acidification processes of the rhizosphere may have caused the decrease in soil pH. In our experiment, the nitrate formation was, in general, higher after the application of organic amendments. Therefore, this process could also have favoured the lowering of pH in comparison to the control soil. At the end of the experiment, no significant differences in soil pH were found among the treatments used. This result was also observed by other authors following the application of organic amendments to calcareous vineyard soils. They attributed this to the intrinsic buffer effect of this soil type [12,21,22].
After the addition of SGM or DC, a significant increase in the oxidisable OC was observed in these soils, with a subsequent stabilisation throughout the experiment ( Figure 3B). This initial increase was higher in the DC-treated soils, in accordance with the higher total OC content of this composted material (Table 1). During the first three months of the experiment, the oxidisable OC content was significantly reduced in most of the amended soils, especially in the case of soils with DC, probably due to organic matter mineralization. Although the organic matter of the wastes undergoes stabilisation during composting, a stable compost can contain a significant proportion of readily available organic compounds, such as microbially-derived hydrolysable sugars [37]. In the amended soils, the content of oxidisable OC was higher in the plots with trellised vines than in the plots with bush vines. This could be due to the OM present in vine root exudates, formed mostly by soluble organic compounds produced during the vegetative development of the plant. Stockmann et al. [38] indicated that root inputs to soil can represent 5%-33% of daily photoassimilate production. This result agrees with the higher grape production in the plots of trellised vines ( Figure 5) and with the possible greater root distribution in the upper soil horizons due to the drip irrigation used in this vine-training system. At the end of the experiment, all the amended soils of the plots where the variety Monastrell was grown as bush vines had oxidisable OC contents similar to those of the control soils. However, in the plots with trellised vines, the final oxidisable OC contents of the soils with the organic amendments were 2-3 times higher than those of the control soil. This result is of interest as the SOM content in the vineyards of this area is very low (<1.5%) [39]. Increases in SOM due to the addition of compost or animal manure have also been reported by other authors [12,13,21,22]. soils in semi-arid areas of southern Tunisia, considering this result as beneficial with regard to increasing the availability of nutrients to plants. These authors indicated that the oxidation of NH4 + from N fertilisers (to produce nitrate), the higher root uptake of cations with respect to anions and the acidification processes of the rhizosphere may have caused the decrease in soil pH. In our experiment, the nitrate formation was, in general, higher after the application of organic amendments. Therefore, this process could also have favoured the lowering of pH in comparison to the control soil. At the end of the experiment, no significant differences in soil pH were found among the treatments used. This result was also observed by other authors following the application of organic amendments to calcareous vineyard soils. They attributed this to the intrinsic buffer effect of this soil type [12,21,22].  After the addition of SGM or DC, a significant increase in the oxidisable OC was observed in these soils, with a subsequent stabilisation throughout the experiment ( Figure  3B). This initial increase was higher in the DC-treated soils, in accordance with the higher total OC content of this composted material (Table 1). During the first three months of the experiment, the oxidisable OC content was significantly reduced in most of the amended soils, especially in the case of soils with DC, probably due to organic matter mineralization. Although the organic matter of the wastes undergoes stabilisation during composting, a stable compost can contain a significant proportion of readily available organic compounds, such as microbially-derived hydrolysable sugars [37]. In the amended soils, the content of oxidisable OC was higher in the plots with trellised vines than in the plots with bush vines. This could be due to the OM present in vine root exudates, formed mostly by soluble organic compounds produced during the vegetative development of the plant. Stockmann et al. [38] indicated that root inputs to soil can represent 5%-33% of daily photoassimilate production. This result agrees with the higher grape production in the plots  Regarding the evolution of the soil organic N content, in all plots, the application of SGM or DC to the soil produced an initial increase, with respect to the control soil ( Figure 4A). In general, the amended soils had organic N contents higher than those of the control soil throughout the experiment, especially in the case of the plots with trellised vines. In the plots with non-trellised Monastrell and trellised Cabernet Sauvignon vines, the two organic amendments produced a similar increase in the final concentration of this nutrient with respect to the control, whereas the final organic N level was significantly higher for the DC treatment in the Monastrell vineyard with trellising. Other authors also observed an increase in the organic N fraction with the addition of organic amendments to the soil in vineyards [12,13,[21][22][23].  The evolution of the NH 4 + -N concentration was similar in all plots ( Figure 4B). The initial values were significantly higher in the amended soils. Later, in most soils, the NH 4 + -N levels had increased on 17 June 2014 and 13 January 2015, when the increase in temperature (in summer) had begun and after the autumn rains, respectively (Figure 1). Morlat and Chaussod [12] and Bustamante et al. [21] also reported that the N mineralisation in vineyard soils can be positively affected by rainfall and temperature. Moreover, the soil tillage carried out before the soil sampling on 13 January 2015 could have contributed to the increased NH 4 + -N content (Figures 1 and 4B), as Steenwerth and Belina [40] also observed in a Mediterranean vineyard agroecosystem with conventional tillage. At the end of the experiment, no great differences in the final values of this parameter were found among the treatments. The NH 4 + -N contents were highest in the SGM soils, probably due to the lower degree of stabilisation of this organic amendment, favouring the mineralisation of organic N [21]. It is also noteworthy that in all plots the NH 4 + -N content was low (0.8-16.7 mg NH 4 + -N kg −1 soil), which is usual in Mediterranean areas where the mineralisation and nitrification rates are high [41]. Therefore, the pool of soluble mineral N in vineyards is dominated by NO 3 − -N during most of the growing season, as observed in this experiment ( Figure 4C) and by Christou et al. [42] in a study on the N dynamic in a Mediterranean vineyard soil amended with poultry manure compost.
In all plots, the nitrate contents were higher in the amended soils than in the control soil ( Figure 4C). This behaviour persisted, in general, throughout the experiment, except at the end when no significant differences due to the treatment were found in the NO 3 − -N concentration. Moreover, in all soils, there was a significant reduction in the NO 3 − -N content at the end of the experiment. This could be due mainly to the leaching of nitrate into deeper layers after the autumn rains (Figure 1). Figure 5 shows the grape yield (kg per vine) for each treatment in the three studied plots. In all plots, it was higher in the treatments with organic amendments than in the control. It is also noteworthy that the yield was higher in the plots with trellised vines, especially in the case of Cabernet Sauvignon, than in the plot with bush vines. This could be due to the rain-fed conditions of the plot with non-trellised vines, which-unlike the plots with trellised vines-did not have drip irrigation. Romero et al. [43] also observed that severe water stress in the vine decreases its fine root development, water uptake and grape production per plant.

Effect of the Organic Fertiliser Type on Greenhouse Gas Emissions
In all plots, the N 2 O flux peaked 7-15 days after the start of the experiment ( Figure 6A). In the control (where organic amendment was not applied), this could be due to the increase in temperature registered at the beginning of the experiment. Some authors have found a positive correlation among temperature, microbial activity and N 2 O fluxes [44,45]. In addition to the impact of an increase in temperature, in the amended soils, this initial growth in N 2 O emissions could be associated with an enhanced availability of labile, Cenriched substrates derived from the added organic matter. These substrates are degraded by soil microorganisms, reducing temporarily the soil O 2 supply and, therefore, promoting anaerobic microsites that are more suitable for denitrification processes and thus hot-spots for the production of N 2 O [46]. Subsequently, the N 2 O flux declined, reaching its lowest values during the dry summer period in most cases, especially in the plots with trellised vines. Beare et al. [47] also observed that the production of N 2 O was strongly reduced during drier periods of the crop season, because of the lower soil microbiological activity. After this dry period, a pulse in the N 2 O flux was observed, probably due to the recovery of the activity of soil microorganisms because of the soil rewetting associated with the September and November rainfall (Figure 1). This increase in the soil biological activity is produced generally by the availability of C and N substrates accumulated in the soil during dry periods, when the soil microorganisms have died [14]. Despite the autumn rains, the soil moisture was not high enough to promote, in combination with the available C and N substrates, highly denitrifying conditions in soil micropores. Therefore, the aerobic oxidation of NH 4 + through nitrification would have been the main process behind the production of N 2 O measured at this stage of the experimental period. Galbally et al. [48] found that in semi-arid conditions the low pulses of N 2 O emissions produced after rainfall events could come from the activity of nitrifying aerobic bacteria. After the soil wetting cycle, the N 2 O emissions diminished until the end of the experiment. It is also noteworthy that the emissions of N 2 O were greater in the plot with bush vines (plot 1) than in plots 2 and 3 with trellised vines, probably due to the synergistic effect of reduced plant growth and lower N use efficiency in the former [49]. This is supported by the lower grape production per plant found in plot 1 ( Figure 5). However, the N 2 O emissions in all plots during the studied growing season were very low compared to those found by Calleja-Cervantes et al. [13] in a vine system with a Mediterranean climate, during a period of 115 days after the addition of different composted materials. These authors found that the soils receiving the organic amendments had maximum N 2 O fluxes between 0.4 and 1.4 mg In the case of the CH 4 emissions of the soil, negative CH 4 fluxes were recorded in all plots during most of the experiment ( Figure 6B). Shvaleva et al. [16] also observed that Mediterranean dry ecosystems can represent a significant sink for atmospheric CH 4 . The low water content of these soils could increase the rate of oxidation of this greenhouse gas, probably due to the higher content of oxygen in the soil porous spaces, which also allows for greater diffusion of CH 4 down the soil profile [50]. The diffusion of CH 4 and O 2 , the main substrates for the activity of methanotrophic bacteria, was also favoured by the coarse texture of our soils [51]. In addition, the mineral N concentration of the studied soils was compatible with the growth of methanotrophic bacteria, since high nitrification rates can produce toxic levels of N 2 O and NH 2 OH that affect CH 4 -consuming bacteria [52]. Moreover, high NH 4 + contents in the soil can reduce CH 4 oxidation, due to the methanotrophic oxidation of NH 4 + being greater than that of CH 4 [52]. Therefore, the low inorganic N content of these soils was probably the main determinant of their function as CH 4 sinks. With regard to soil respiration, the CO 2 flux emitted from the soil varied enormously throughout the experimental period, in all plots ( Figure 6C). The range of values of these fluxes was 7-1518 mg C-CO 2 m −2 d −1 , the highest values being recorded in the amended soils. In general, the emission of this greenhouse gas was highest during the first 42 days of the experiment, possibly due to the addition of the organic amendments and to the tillage carried out in early April. An increase in CO 2 emission with the addition of organic amendments is commonly attributed to an increase in the metabolic activity of the microorganisms present in the soil, due to the availability of easily degradable compounds from these amendments [53]. In this way, increases in the emission of CO 2 from agricultural soils after the addition of organic amendments have been previously reported by different authors [54,55]. In general, in the vine systems studied here, the SGM treatment generated higher emissions of CO 2 than the compost used. This may well be related to the decomposition rate of the organic amendments [54], since SGM has an OM content that is more labile and less stabilised than that of the compost. Soil tillage can also contribute to the CO 2 flux emitted from the soil, possibly due to greater soil aeration and the breakdown of soil macroaggregates, which leaves previously protected labile OM prone to decomposition [17]. The CO 2 emissions were lowest in the dry summer period (July and August), when the water content of the soil was low, even in the plots with trellised vines, where irrigation was carried out. Shvaleva et al. [16] also observed this reduction in CO 2 emissions from the soil in the summer, even in irrigated plots, in a study of the influence of the rainfall and irrigation on the GHG emissions in oak forests in southern Portugal. It is also of note that the emissions of CO 2 were greater in the plots with trellised vines, especially in the case of Cabernet Sauvignon (plot 3). This is probably related to the higher grape production of this variety ( Figure 5), with greater root exudation of labile C resulting in increased CO 2 emissions during the vegetative development of the plant [48]. However, the CO 2 emissions in the present experiment are within the range of values observed by Lardo et al. [8] in a study of the dynamics of the total edaphic respiration over a period of three years in vineyard soils of southern Italy (213-11,978 mg C-CO 2 m −2 d −1 ). Figure 7A-C shows the cumulative N 2 O, CH 4 and CO 2 emissions, respectively, during the vine growing season, as affected by the different treatments. In plot 1 (non-trellised Monastrell vines), the DC treatment gave the highest accumulated N 2 O emissions at the end of the experiment (297 days), whereas no significant differences were observed in these emissions between the SGM-amended and control soils. In the plot with trellised Monastrell vines, no statistically significant differences in the cumulative N 2 O emissions were observed between the compost treatment and the control, the values being highest in the SGM soil ( Figure 7A). In plot 3, with trellised Cabernet Sauvignon vines, the cumulative N 2 O emissions were lowest in the control soil and highest in the DC soil, while the SGM soil had intermediate values. The cumulative emissions of N 2 O were greater in the plot with bush vines (plot 1) than in those with trellised vines (plots 2 and 3). However, the cumulative N 2 O emissions, in all plots, were very low compared to those found by Aguilera et al. [14] in vine systems in Mediterranean climates (0.4 ± 0.2 kg N-N 2 O ha −1 year −1 ), in a meta-analysis of the emissions of this greenhouse gas in different Mediterranean agroecosystems. These authors also found that among the different types of crop studied (vineyard, horticultural, cereals and legumes) the vineyards gave the lowest N 2 O emissions, associated with the low inputs of N and water of these agricultural systems.
Significant differences in the accumulated emissions of CH 4 were observed among the treatments tested in the three plots of this study ( Figure 7B). In the plot with non-trellised Monastrell vines, the CH 4 consumption was lowest in the SGM soil, while the control and DC soils did not differ significantly in their negative emissions of CH 4 . However, the SGM and DC soils showed the highest capacity for CH 4 oxidation, in comparison to the control soil, in plot 2 (trellised Monastrell vineyard). Moreover, in the plot with trellised Cabernet Sauvignon vines the CH 4 consumption was highest in the control soil, while it was statistically similar in the DC and SGM soils.
Regarding the accumulated soil emissions of CO 2 , in the plot with non-trellised Monastrell vines, no significant differences were found among the treatments tested ( Figure 7C). However, in the plot with the trellised Monastrell vines, the SGM and control treatments gave the highest and lowest accumulated emissions of CO 2 , respectively. Furthermore, in the plot with the trellised Cabernet Sauvignon vines there were significant differences for both treatments, with respect to the control: the emission of CO 2 was highest for the SGM treatment, followed by DC and then the control. In general, in the vine systems studied here, the SGM treatment generated higher cumulative emissions of CO 2 than the compost used. This may well be related to the decomposition rate of the organic amendments used [54], since SGM has an OM content that is more labile and less stabilised than that of the compost.
Agronomy 2021, 11, x. https://doi.org/10.3390/xxxxx www.mdpi.com/journal/agronomy Figure 7A-C shows the cumulative N2O, CH4 and CO2 emissions, respectively, during the vine growing season, as affected by the different treatments. In plot 1 (non-trellised Monastrell vines), the DC treatment gave the highest accumulated N2O emissions at the end of the experiment (297 days), whereas no significant differences were observed in these emissions between the SGMamended and control soils. In the plot with trellised Monastrell vines, no statistically significant differences in the cumulative N2O emissions were observed between the compost treatment and the control, the values being highest in the SGM soil ( Figure 7A). In plot 3, with trellised Cabernet Sauvignon vines, the cumulative N2O emissions were lowest in the control soil and highest in the DC soil, while the SGM soil had intermediate values. The cumulative emissions of N2O were greater in the plot with bush vines (plot 1) than in those with trellised vines (plots 2 and 3). However, the cumulative N2O emissions, in all plots, were very low compared to those found by Aguilera et al. [14] in vine systems in Mediterranean climates (0.4 ± 0.2 kg N-N2O ha −1 year −1 ), in a meta-analysis of the emissions of this greenhouse gas in different Mediterranean agroecosystems. These authors also found that among the different types of crop studied (vineyard, horticultural, cereals and legumes) the vineyards gave the lowest N2O emissions, associated with the low inputs of N and water of these agricultural systems. Significant differences in the accumulated emissions of CH4 were observed among the treatments tested in the three plots of this study ( Figure 7B). In the plot with nontrellised Monastrell vines, the CH4 consumption was lowest in the SGM soil, while the control and DC soils did not differ significantly in their negative emissions of CH4. However, the SGM and DC soils showed the highest capacity for CH4 oxidation, in comparison to the control soil, in plot 2 (trellised Monastrell vineyard). Moreover, in the plot with trellised Cabernet Sauvignon vines the CH4 consumption was highest in the control soil, while it was statistically similar in the DC and SGM soils.
Regarding the accumulated soil emissions of CO2, in the plot with non-trellised Monastrell vines, no significant differences were found among the treatments tested ( Figure  7C). However, in the plot with the trellised Monastrell vines, the SGM and control treatments gave the highest and lowest accumulated emissions of CO2, respectively. Furthermore, in the plot with the trellised Cabernet Sauvignon vines there were significant differences for both treatments, with respect to the control: the emission of CO2 was highest for the SGM treatment, followed by DC and then the control. In general, in the vine systems studied here, the SGM treatment generated higher cumulative emissions of CO2 than the compost used. This may well be related to the decomposition rate of the organic amendments used [54], since SGM has an OM content that is more labile and less stabilised than that of the compost. Figure 8 shows the plot of the FCs extracted by FA, for the soil parameters and GHG emissions studied (n = 8). This statistical analysis had a measure of sampling adequacy (Kaiser-Meyer-Olkin) above 0.5, the Bartlett sphericity test gave a P value of 0.000 and all variables had an extraction value >0.5. These data indicate that the model proposed was suitable. This model grouped the studied parameters into two FCs, with a cumulative explained variance >70%: FC1 explained 48.9% and FC2 22.1%. In FC1, the following parameters were grouped: organic N, oxidisable OC, pH, N2O-N emission and NH4 + -N. All  Figure 8 shows the plot of the FCs extracted by FA, for the soil parameters and GHG emissions studied (n = 8). This statistical analysis had a measure of sampling adequacy (Kaiser-Meyer-Olkin) above 0.5, the Bartlett sphericity test gave a P value of 0.000 and all variables had an extraction value >0.5. These data indicate that the model proposed was suitable. This model grouped the studied parameters into two FCs, with a cumulative explained variance >70%: FC1 explained 48.9% and FC2 22.1%. In FC1, the following parameters were grouped: organic N, oxidisable OC, pH, N 2 O-N emission and NH 4 + -N. All these variables were positively correlated among themselves, except pH, indicating that the increase in soil organic C and N caused by the addition of the organic amendments contributed to the reduction of soil pH, as was also found by Hannachi et al. [36] in alkaline cultivated soils of southern Tunisia. In addition, the NH 4 + -N content and N 2 O-N emission were positively correlated, possibly because in this agroecosystem the nitrification processes contributed to a greater extent to the emissions of N 2 O than the denitrification processes. Other authors also found in their studies that N 2 O emissions could be explained better by the soil NH 4 + content than by the NO 3 − content, due to the nitrification processes [15,45]. In FC2, the following parameters were grouped and positively correlated: NO 3 − -N, CH 4 -C emission and CO 2 -C emission. A positive correlation between CO 2 and CH 4 emissions was also observed by Shvaleva et al. [16] in an experiment carried out to study the influence of rainfall and irrigation on the GHG emissions in oak forests in southern Portugal. These authors stated that this correlation was due to the close relationship between the cycles of C and N, which involve microorganisms that directly contribute to the fluxes of CO 2 , N 2 O and CH 4 in the soil. Moreover, the positive correlation between the NO 3 − -N content and CH 4 emissions could be related to the inhibition of CH 4 oxidation by soil microorganisms when the concentration of nitrate in the soil is high, producing higher CH 4 emissions, as mentioned in the section Effect of the organic fertiliser type on greenhouse gas emissions. tween the cycles of C and N, which involve microorganisms that directly contribute to the fluxes of CO2, N2O and CH4 in the soil. Moreover, the positive correlation between the NO3 − -N content and CH4 emissions could be related to the inhibition of CH4 oxidation by soil microorganisms when the concentration of nitrate in the soil is high, producing higher CH4 emissions, as mentioned in the section Effect of the organic fertiliser type on greenhouse gas emissions.

Conclusions
This work has evaluated the short-term effect of the organic fertiliser type on soil fertility, grape production and greenhouse gas emissions in semi-arid vineyard agroecosystems with different agricultural practices. The use of organic amendments increased the grape yield and improved the fertility of the soil, especially in the case of the dripirrigated trellising vineyards. Regarding the environmental aspects, in general, the use of manure generated higher emissions of CO2 than the compost used. Moreover, higher grape production enhanced the CO2 emissions, as observed in the case of trellised Cabernet Sauvignon vines. Nitrification was the main soil microbiological process leading to the production of N2O in the studied soils, mostly due to the low soil moisture. The N2O emissions were greatest in the plot with bush vines, where the grape yield was lowest. In addition, the low water and mineral N contents of the soils and their coarse texture contribute to the relevance of these vineyard agrosystems as important sinks for atmospheric

Conclusions
This work has evaluated the short-term effect of the organic fertiliser type on soil fertility, grape production and greenhouse gas emissions in semi-arid vineyard agroecosystems with different agricultural practices. The use of organic amendments increased the grape yield and improved the fertility of the soil, especially in the case of the drip-irrigated trellising vineyards. Regarding the environmental aspects, in general, the use of manure generated higher emissions of CO 2 than the compost used. Moreover, higher grape production enhanced the CO 2 emissions, as observed in the case of trellised Cabernet Sauvignon vines. Nitrification was the main soil microbiological process leading to the production of N 2 O in the studied soils, mostly due to the low soil moisture. The N 2 O emissions were greatest in the plot with bush vines, where the grape yield was lowest. In addition, the low water and mineral N contents of the soils and their coarse texture contribute to the relevance of these vineyard agrosystems as important sinks for atmospheric CH 4 . Considering the agronomic and environmental aspects in an integrated manner, the optimal vineyard management to ensure high grape yields and to improve the soil fertility with low greenhouse gas emissions were the use of compost in the drip-irrigated trellising vineyard with the grape variety Monastrell.
Overall, this work indicates that from a socio-environmental perspective, the use of organic amendments derived from the agro-food industry helps to close nutrient and energy loops, thus leading to a lower environmental impact of the food production sector within a regionally-based concept of the circular economy. The use of organic amendments should be analysed more deeply and generalised towards the sustainability of agro-food systems, particularly in semi-arid regions that are highly vulnerable to the impacts of regional and global climate changes. Therefore, more and longer studies than the present work are required to determine the active contributions of different agricultural variables to the mitigation of climate change in the ecosystems of Mediterranean climates.