Potential Role of Fertilizer Sources and Soil Tillage Practices to Mitigate Soil CO2 Emissions in Mediterranean Potato Production Systems

Agricultural practices should be approached with environmental-friendly strategies, able to restore soil organic matter and reduce the greenhouse gas emissions. The main objective of this study is to evaluate the environmental benefits, in terms of CO2 emissions and carbon balance, of some agricultural practices for potato cultivation. A randomized complete block design was adopted where the treatments were: (a) tillage systems (plowing; subsoiler and spading); (b) fertilizer sources (mineral and organic). All treatments were replicated three times. Potato yield and its carbon content, soil CO2 emissions, temperature, and volumetric water content were measured. The CO2 emissions were higher in organic than in mineral fertilizer (0.60 and vs. 0.77 g m−2 h−1, respectively), while they were low in spading compared to the other soil tillage (0.64 vs. 0.72 g m−2 h−1, respectively). Carbon input was the highest in plowing and organic fertilizer 4.76 and 5.59 Mg C ha−1, respectively. The input/output ratio of carbon varied according to the main treatments. The findings suggest that spading tillage and organic fertilizer might result in environmental and agronomical benefits, further research should be performed to evaluate to possibility to extend the results to other environments and crops.


Introduction
Nowadays, the adoption of farming practices that increase soil carbon (C) stocks are strongly recommended to improve the sustainability of the agricultural systems in all environments. Although intensive agricultural management has contributed to the economic and social development of modern society, it has also led to a progressive environmental degradation by increasing the greenhouse gas (GHG) emissions and reducing soil quality [? ]. In general, agriculture is considered as one of the most important human activities that produce significant amount of GHG emissions, and the emissions released from soil into the atmosphere represent a large amount of carbon dioxide (CO 2 ) [? ].
Soil represents the biggest terrestrial pool of C, and the soil tillage approach, mainly used for seedbed preparation and weeding, is the common practice adopted since the human became a farmer. Soil acts as a source of nutrients and water, but the modification of soil aggregation and structure due to the modern agricultural practices may increase considerably the surface runoff and losses of nutrients and soil [? ]. According to Abdalla et al. [? ], conventional soil tillage, based on deep and frequent

Study Site and Soil Characteristics
The research was conducted at the experimental farm "Nello Lupori" of the University of Tuscia located in central Italy (latitude 45 • 25 , longitude 12 • 04 and altitude 310 m a.s.l.) during the 2015 and 2016 cropping seasons. The climate of the area is typically Mediterranean, with warm and wet winters and hot and dry summers. The average temperature is 14.5 • C, with minimum temperatures just below 0 • C in February and maximum temperature above 35 • C in July. Annual rainfall is on average 752 mm (last 30 years), mainly spread from September to May (569 mm). The soil of the experimental area is of volcanic origin and classified as Typic Xerofluvent [? ]. The particle size distribution of the soil in the surface horizon (0-30 cm depth) were 76.3% sand, 13.3% silt, and 10.4% clay and indicate the soil as loamy sand. The soil had 0.97% and 0.12% of total organic C and N content, respectively, and pH (H 2 O) 6.9.

Experimental Design
A 2-year crop rotation durum wheat (Triticum durum Desf.)-Potato (Solanum tuberosum L.), typical of the study area, was established in 2013 in order to compare soil tillage practices and fertilizer sources. The treatments consisted of (a) three tillage systems (conventional tillage based on mould-board plough; reduced tillage based on subsoiler; reduced tillage based on the spading machine), and (b) two fertilizer sources (mineral fertilization as performed in conventional farming and organic fertilization by means of municipal organic waste). The treatments were replicated three times according to a randomized complete block design. Considering that both crops in rotation were simultaneously cropped, the experimental field included 36 plots (two crops × three soil tillage × two fertilizer source × three blocks). Each experimental plot was 60 m 2 (6 × 10 m) and they were separated by 3 m wide alleys to allow for equipment operations.

Field Setup and Crop Management
Every year in April, before planting potato tubers, the soil was tilled according to the selected treatments. All tillage operations were performed to a depth of 20 cm and then followed by a seedbed preparation with a disk harrowing (10 cm of soil depth) carried out just prior to potato planting. Potato cv. Monalisa was planted by using seed pieces of tuber previously cut. Potato tuber pieces were hand planted in open furrows on 13 April 2015 and 31 March 2016, respectively, at the density of 7.14 plants m −2 with the spacing of plant rows and plant in rows of 0.70 m and 0.20 m, respectively. The same plant geometry was adopted in all experimental treatments. In the mineral fertilizer treatments, fertilizers were applied twice; the first time, at potato planting, 100 kg ha −1 of P 2 O 5 as triple perphosphate, 50 kg ha −1 of K 2 O as potassium sulphate, and 100 kg ha −1 of N as urea, while the second fertilizer rate was applied 6 weeks after planting, at hilling stage, using the equivalent of 70 kg ha −1 of N as ammonium nitrate. For organic fertilization treatments, an amount of 18 Mg ha −1 of mature organic waste was totally applied at planting before seedbed preparation. In both fertilization source treatments, potato received identical amounts of total N. The rate of fertilization was calculated based on the demand of 170 kg ha −1 of N according to the common practices adopted by the farmers in the study area [? ] and considering the mineralization rate. Hilling up was performed a month after the planting, using a tractor-mounted cultivator. Irrigation water was supplied when soil moisture reached 65% of the soil field capacity in each experimental year and stopped 2 weeks before tuber harvesting. A sprinkler irrigation system was used to irrigate the potato crop five times in 2015 and four in 2016, respectively. Potato tubers were harvested on 28 July 2015 and 4 August 2016, when about 70% of haulms were dry.

Data Collection and Analysis
Weather data characteristics (minimum, maximum and mean temperature, and precipitations) for the entire growing seasons of potato in both 2015 and 2016 years were measured by a meteorological station located at 500 m from the experimental site. Fluxes of CO 2 emissions from all experimental treatments were measured during both potato growing seasons, from April to July, using a portable dynamic closed-chamber infrared gas analyzer system [? ]. A non-steady-state through-flow chamber (SRC-1, PP Systems, Stofold, UK) with a volume of 1334 cm 3 and a cover area of 78.5 cm 2 was used to perform soil CO 2 emission measurements. The chamber had only one opening to the soil and the soil CO 2 emission readings were performed placing the chamber on a small permanent PVC frame carefully placed in a central area of each plot in a potato row between two consecutive plants [? ? ]. The area inside the PVC collar was kept weed-free by clipping weeds whenever necessary. Measurements were conducted weekly at the same time (between 8:00 a.m. and 10:00 a.m.) in order to reduce diurnal variation by means of a sensitive infrared gas analyzer instrument (EGM-4, PP Systems, Stofold, UK) [? ]. Each measurement took about 180 s and varied based on the respiration rate of the soil. The amount of soil CO 2 emissions was estimated by fitting a quadratic equation to the relationship between the increasing CO 2 concentration and elapsed time. The values obtained were considered as the net carbon mineralization since the autotrophic component was included. The estimation was determined considering two consecutive measurements in a linear interpolation and their integration over time (trapezoid rule) as previously adopted by Mancinelli et al. [? ].
Soil temperature at 0-5 cm depth and soil volume water content at 0-30 cm depth was measured simultaneously to soil CO 2 emissions near to the chamber by using STP1 soil temperature Probe connected to EGM-4 instrument and TDR 300 Soil Moisture Meter (Spectrum Technologies, Inc. Plainfield, IL, USA), respectively.
At potato harvesting, five soil cores (0-20 cm depth) were collected from each plot and pooled together for the soil analysis. Soil samples were taken from the central inter-row of each plot, sieved (<2 mm) and stored at 4 • C until their use for total organic carbon (TOC) and total nitrogen (TN) determination. The TOC and TN contents were determined using an elementar analyzer (Thermo Soil NC-Flash EX1112).
Tuber yield and aboveground biomass of potato were sampled in 2 m 2 of a central area of each plot and crop physiological maturity. Tubers were classified as marketable (graded by tuber diameter between 40 and 75 mm) and unmarketable tubers (cull tubers with a diameter <40 mm). Samples were oven-dried at 70 • C until constant weight and finely ground, in order to determine the carbon (C) content by means of an elementar analyzer (Thermo Soil NC-Flash EX1112).
The ration of carbon input and carbon output (C input/C output) was determined considering the carbon produced by straw and tubers of potato crop (C input) and the carbon content in the CO 2 emissions during the potato growing season (C output). The ratio offers an evaluation of the direction of the carbon cycle. Values higher, equal to, or lower than 1 mean increasing, stable or decreasing soil carbon content in the system, respectively. The ratio between the cumulative carbon dioxide emissions and the tuber yield was used to evaluate the amount of CO 2 emissions per unit of potato yield. The lower values mean lower soil CO 2 emission per unit of tuber yield [? ].

Statistical Analysis
Analysis of variance was performed for the data on soil tillage and fertilization source obtained from the experimental fields. The JMP statistical software package 4.0 [? ] was used for the statistical analysis. The ANOVA was conducted for the 2 year period applying a randomized complete block design with three blocks. A two-way factorial experimental design was performed for measured characteristics where the soil tillage and fertilization source were the treatments, and the year was considered as a repeated measure [? ]. Means were compared adopting the Fisher's protected least significant difference (LSD) at p < 0.05. The soil CO 2 flux, soil temperature and soil volumetric water content are presented as means of the considered treatment ± standard error (SE). A second-order polynomial regression model of soil CO 2 emissions by soil temperature and soil volume water content were performed separately for each soil tillage and fertilization source treatments.

Weather Conditions during Potato Growing Seasons
The potato growing period was from March to August and lasted for 106 days in 2015 and 126 days in 2016, respectively. The weather variables as air temperature and rainfall were recorded during the whole potato growing period in both experimental seasons (Figure ??). The cumulated precipitation was higher in 2016 than 2015 (181.0 vs. 147.8 mm, respectively), even if they tended to be distributed differently between the potato growing seasons. Indeed, the precipitation in 2015 growing seasons was mainly concentrated in the first month after potato sowing (about 60% of the total precipitation), while in 2016 the precipitation was mainly distributed in May-June (about 70% of the total precipitation). As expected, minimum and maximum temperatures tended to gradually increase from March to August when the air temperature reached the highest values of the year (35.3 • C and 32.9 • C in 2015 and 2016, respectively). The mean air temperature was higher in 2015 than in 2016 potato growing season (on average 19.0 • C vs. 18.2 • C, respectively).

Weather Conditions during Potato Growing Seasons
The potato growing period was from March to August and lasted for 106 days in 2015 and 126 days in 2016, respectively. The weather variables as air temperature and rainfall were recorded during the whole potato growing period in both experimental seasons ( Figure 1). The cumulated precipitation was higher in 2016 than 2015 (181.0 vs. 147.8 mm, respectively), even if they tended to be distributed differently between the potato growing seasons. Indeed, the precipitation in 2015 growing seasons was mainly concentrated in the first month after potato sowing (about 60% of the total precipitation), while in 2016 the precipitation was mainly distributed in May-June (about 70% of the total precipitation). As expected, minimum and maximum temperatures tended to gradually increase from March to August when the air temperature reached the highest values of the year (35.3 °C and 32.9 °C in 2015 and 2016, respectively). The mean air temperature was higher in 2015 than in 2016 potato growing season (on average 19.0 °C vs. 18.2 °C, respectively).

Effect of Fertilizer Sources
The flux of soil CO2 emissions, soil temperature and soil volume water content during the potato growing seasons under organic and mineral fertilizer sources are reported in Figure 2. The soil CO2 emissions showed different values and trends between the 2015 and 2016 growing seasons. In 2015, the CO2 emissions trend slightly increased in both fertilizer sources, reaching the highest value in May (on average 0.71 g m −2 h −1 ), then it gradually decreased showing the lowest value of CO2 emissions just before potato harvesting (on average 0.38 g m −2 h −1 ). In 2016 the CO2 emissions rapidly increase after potato sowing reaching a peak after one month in May, with an average value between the fertilizer sources of 1.01 g m −2 h −1 , then the flux of CO2 emissions tended to decrease slowly until the potato harvesting (on average 0.42 g m −2 h −1 , Figure 2). Although the trend was similar in mineral and organic fertilizer sources, the soil CO2 emissions were generally higher in organic than in mineral fertilizer (on average 0.

Weather Conditions during Potato Growing Seasons
The potato growing period was from March to August and lasted for 106 days in 2015 and 126 days in 2016, respectively. The weather variables as air temperature and rainfall were recorded during the whole potato growing period in both experimental seasons ( Figure 1). The cumulated precipitation was higher in 2016 than 2015 (181.0 vs. 147.8 mm, respectively), even if they tended to be distributed differently between the potato growing seasons. Indeed, the precipitation in 2015 growing seasons was mainly concentrated in the first month after potato sowing (about 60% of the total precipitation), while in 2016 the precipitation was mainly distributed in May-June (about 70% of the total precipitation). As expected, minimum and maximum temperatures tended to gradually increase from March to August when the air temperature reached the highest values of the year (35.3 °C and 32.9 °C in 2015 and 2016, respectively). The mean air temperature was higher in 2015 than in 2016 potato growing season (on average 19.0 °C vs. 18.2 °C, respectively).

Effect of Fertilizer Sources
The flux of soil CO2 emissions, soil temperature and soil volume water content during the potato growing seasons under organic and mineral fertilizer sources are reported in Figure 2. The soil CO2 emissions showed different values and trends between the 2015 and 2016 growing seasons. In 2015, the CO2 emissions trend slightly increased in both fertilizer sources, reaching the highest value in May (on average 0.71 g m −2 h −1 ), then it gradually decreased showing the lowest value of CO2 emissions just before potato harvesting (on average 0.38 g m −2 h −1 ). In 2016 the CO2 emissions rapidly increase after potato sowing reaching a peak after one month in May, with an average value between the fertilizer sources of 1.01 g m −2 h −1 , then the flux of CO2 emissions tended to decrease slowly until the potato harvesting (on average 0.42 g m −2 h −1 , Figure 2). Although the trend was similar in mineral and organic fertilizer sources, the soil CO2 emissions were generally higher in organic than in mineral fertilizer (on average 0.

Effect of Fertilizer Sources
The flux of soil CO 2 emissions, soil temperature and soil volume water content during the potato growing seasons under organic and mineral fertilizer sources are reported in Figure ??. The soil CO 2 emissions showed different values and trends between the 2015 and 2016 growing seasons. In 2015, the CO 2 emissions trend slightly increased in both fertilizer sources, reaching the highest value in May (on average 0.71 g m −2 h −1 ), then it gradually decreased showing the lowest value of CO 2 emissions just before potato harvesting (on average 0.38 g m −2 h −1 ). In 2016 the CO 2 emissions rapidly increase after potato sowing reaching a peak after one month in May, with an average value between the fertilizer sources of 1.01 g m −2 h −1 , then the flux of CO 2 emissions tended to decrease slowly until the potato harvesting (on average 0.42 g m −2 h −1 , Figure ??). Although the trend was similar in mineral and organic fertilizer sources, the soil CO 2 emissions were generally higher in organic than in mineral fertilizer (on average 0.65 and 0.81 vs. 0.54 and 0.72 g m −2 h −1 in 2015 and in 2016, respectively). However, the biggest differences were observed at the beginning of the potato cultivation in 2015 and from the middle potato season in 2016 (Figure ??). Soil temperature followed the weather conditions and tended to increase from the potato planting (on average 15.6 • C and 15.7 • C in 2015 and in 2016, respectively) to harvesting (on average 29.7 • C and 27.6 • C in 2015 and in 2016, respectively). Few differences were observed on soil temperature between organic and mineral fertilizer sources (Figure ??). The regression analyses reported in Figure ??A showed that the temperature dependency of soil CO 2 flux is well described through polynomial regression analysis. The range of values measured over the two years of experimentation, the polynomial regression curve shows the highest level of soil CO 2 emission at optimum soil temperature in the organic fertilizer source (0.80 g m −2 h −1 , R 2 = 0.1926, p < 0.001) compared with mineral fertilizer (0.68 g m −2 h −2 , R 2 = 0.1416, p < 0.001, Figure ??A). Similarly, the volumetric water content of the soil followed the climatic conditions, in particular the precipitation events. In fact, it was high after potato planting in 2015, while the volumetric water content of soil tended to be low at the beginning of potato cultivation and before potato harvesting, but there were peaks in the middle of potato growing seasons (Figure ??). The trend of the soil volumetric water content was similar between the organic and fertilizer sources. A polynomial regression analysis reported in Figure ??B showed the dependency of soil CO 2 flux on soil volumetric water content in all fertilizer treatments. The curve, obtained with the values measured over the two years of experiments, showed the highest level of soil CO 2 emissions at optimum soil volume content in the organic fertilizer source (0.82 g m −2 h −1 , R 2 = 0.1462, p < 0.001) with respect to mineral fertilizer source (0.76 g m −2 h −1 , R 2 = 0.1468, p < 0.001, Figure ??B). . Few differences were observed on soil temperature between organic and mineral fertilizer sources (Figure 2). The regression analyses reported in Figure 3A showed that the temperature dependency of soil CO2 flux is well described through polynomial regression analysis. The range of values measured over the two years of experimentation, the polynomial regression curve shows the highest level of soil CO2 emission at optimum soil temperature in the organic fertilizer source (0.80 g m −2 h −1 , R 2 = 0.1926, p < 0.001) compared with mineral fertilizer (0.68 g m −2 h −2 , R2 = 0.1416, p < 0.001, Figure 3A). Similarly, the volumetric water content of the soil followed the climatic conditions, in particular the precipitation events. In fact, it was high after potato planting in 2015, while the volumetric water content of soil tended to be low at the beginning of potato cultivation and before potato harvesting, but there were peaks in the middle of potato growing seasons ( Figure  2). The trend of the soil volumetric water content was similar between the organic and fertilizer sources. A polynomial regression analysis reported in Figure 3B showed the dependency of soil CO2 flux on soil volumetric water content in all fertilizer treatments. The curve, obtained with the values measured over the two years of experiments, showed the highest level of soil CO2 emissions at optimum soil volume content in the organic fertilizer source (0.82 g m −2 h −1 , R 2 = 0.1462, p < 0.001) with respect to mineral fertilizer source (0.76 g m −2 h −1 , R 2 = 0.1468, p < 0.001, Figure 3B).

Effect of Soil Tillage Management
In Figure 4, the flux of soil CO2 emissions, temperature, and soil volume water content during the potato growing seasons under various treatments are reported. In both potato growing seasons, the soil CO2 emissions were generally low in spading compared with other soil tillage (on average 0.64 vs. 0.72 g m −2 h −1 , respectively), even if the biggest differences were observed in 2015 (0.51 vs. 0.63 g m −2 h −1 , respectively, Figure 4). In both potato growing seasons, soil temperatures were similar in the tillage treatments, especially in the period from potato sowing to flowering, while few differences on soil temperature were observed before potato harvesting (Figure 4). The regression analyses reported in Figure 3C showed that the soil temperature dependency of soil CO2 flux was well described through a polynomial regression analysis. The polynomial regression curve, obtained using the range of values measured over the two years of experimentation, showed the highest level of soil CO2 emission at optimum soil temperature in the subsoiling tillage (0.79 g m −2 h −1 , R 2 = 0.2398, p < 0.001), intermediate in plowing (0.75 g m −2 h −1 , R 2 = 0.1011, p < 0.001) and the lowest in spading (0.0.72 g m −2 h −1 , R 2 = 0.1722, p < 0.001, Figure 3C).
The volumetric water content of soil was similar among the soil tillage treatments (Figure 4). A polynomial regression analysis, reported in Figure 3D, showed the dependency of soil CO2 flux on soil volumetric water content in all tillage treatments. The curve, obtained with the values measured over the two years of experimentation, showed the highest level of soil CO2 emissions at optimum soil temperature in the subsoiling tillage (0.86 g m −2 h −1 , R 2 = 0.1476, p < 0.001), intermediate in plowing (0.81 g m −2 h −1 , R 2 = 0.2084, p < 0.001) and the lowest in spading (0.79 g m −2 h −1 , R 2 = 0.0951, p < 0.001, Figure 3D).

Effect of Soil Tillage Management
In Figure ??, the flux of soil CO 2 emissions, temperature, and soil volume water content during the potato growing seasons under various treatments are reported. In both potato growing seasons, the soil CO 2 emissions were generally low in spading compared with other soil tillage (on average 0.64 vs. 0.72 g m −2 h −1 , respectively), even if the biggest differences were observed in 2015 (0.51 vs. 0.63 g m −2 h −1 , respectively, Figure ??). In both potato growing seasons, soil temperatures were similar in the tillage treatments, especially in the period from potato sowing to flowering, while few differences on soil temperature were observed before potato harvesting (Figure ??). The regression analyses reported in Figure ??C showed that the soil temperature dependency of soil CO 2 flux was well described through a polynomial regression analysis. The polynomial regression curve, obtained using the range of values measured over the two years of experimentation, showed the highest level of soil CO 2 emission at optimum soil temperature in the subsoiling tillage (0.79 g m

Potato Yield and Agro-Ecosystem Carbon Balance
Marketable potato tuber yield was affected by treatments and significant interactions were observed with years (year x soil tillage and year x fertilizer sources) (

Potato Yield and Agro-Ecosystem Carbon Balance
Marketable potato tuber yield was affected by treatments and significant interactions were observed with years (year x soil tillage and year x fertilizer sources) (Figure ??) The total biomass produced by potato crop at harvesting (straw + tubers), determined as the C input obtained by the system, was affected by year, soil tillage, and fertilizer source as main treatments adopted in this study (Table 1). It was higher in 2016 than in 2015 (4.63 vs. 4.35 Mg C ha −1 , respectively). Among the soil tillage treatments, the C input produced by biomasses was higher in plowing compared to subsoiling and spading (on average 4.76 vs. 4.36 Mg C ha −1 , respectively), while considering the fertilizer source, it was greater in organic than in mineral fertilizer source (5.59 vs. 3.40 Mg C ha −1 , respectively).
The C output in the experiment was represented by the carbon loss with the CO2 emissions during the potato cultivation. Generally, the total C from CO2 was greater in 2016 than in 2015 growing seasons (4.39 vs. 3.08 Mg C ha −1 , respectively). However, the plowing and subsoiling tillage showed similar C output by CO2 emissions (on average 3.88 Mg C ha −1 ), which was greater compared to the C output by CO2 emissions measured under spading treatments (3.45 Mg C ha −1 , Table 1). Moreover, the organic fertilizer source determined higher C output by CO2 than mineral fertilizer source (3.97 vs. 3.51 Mg C ha −1 , respectively).
The C input/C output ratio of potato crop varied according to the main treatments. It was the highest in spading (1.25), intermediate in plowing (1.24), and low in subsoiling (1.12), while between the fertilizer source it was greater under organic compared with mineral fertilizer source (1.41 vs. 0.97, respectively, Table 1). In 2016, the C input/C output ratio was higher than in 2015 potato growing season (1.41 vs. 1.05, respectively).
The cumulative CO2/tuber yield ratio was greater in 2015 than in 2016 (on average 0.80 vs. 0.57, respectively). However, it was larger in subsoiling compared to plowing and spading soil tillage (on average 0.89 vs. 0.58, respectively), while organic fertilizer source showed a greater CO2 emission/yield ratio compared with mineral fertilizer source (0.75 vs. 0.62, respectively, Table 1). The total biomass produced by potato crop at harvesting (straw + tubers), determined as the C input obtained by the system, was affected by year, soil tillage, and fertilizer source as main treatments adopted in this study (Table ??). It was higher in 2016 than in 2015 (4.63 vs. 4.35 Mg C ha −1 , respectively). Among the soil tillage treatments, the C input produced by biomasses was higher in plowing compared to subsoiling and spading (on average 4.76 vs. 4.36 Mg C ha −1 , respectively), while considering the fertilizer source, it was greater in organic than in mineral fertilizer source (5.59 vs. 3.40 Mg C ha −1 , respectively). The C output in the experiment was represented by the carbon loss with the CO 2 emissions during the potato cultivation. Generally, the total C from CO 2 was greater in 2016 than in 2015 growing seasons (4.39 vs. 3.08 Mg C ha −1 , respectively). However, the plowing and subsoiling tillage showed similar C output by CO 2 emissions (on average 3.88 Mg C ha −1 ), which was greater compared to the C output by CO 2 emissions measured under spading treatments (3.45 Mg C ha −1 , Table ??). Moreover, the organic fertilizer source determined higher C output by CO 2 than mineral fertilizer source (3.97 vs. 3.51 Mg C ha −1 , respectively).
The C input/C output ratio of potato crop varied according to the main treatments. It was the highest in spading (1.25), intermediate in plowing (1.24), and low in subsoiling (1.12), while between the fertilizer source it was greater under organic compared with mineral fertilizer source (1.41 vs. 0.97, respectively, Table ??). In 2016, the C input/C output ratio was higher than in 2015 potato growing season (1.41 vs. 1.05, respectively).
The cumulative CO 2 /tuber yield ratio was greater in 2015 than in 2016 (on average 0.80 vs. 0.57, respectively). However, it was larger in subsoiling compared to plowing and spading soil tillage (on average 0.89 vs. 0.58, respectively), while organic fertilizer source showed a greater CO 2 emission/yield ratio compared with mineral fertilizer source (0.75 vs. 0.62, respectively, Table ??).

Discussion
Management practices adopted for agricultural production have a direct implication on soil functions, by modifying the diversity and abundance of soil microorganisms with consequences on the ecosystem services and, therefore, on the sustainability of agro-ecosystems [? ]. As any management practice shows effects on crop production, it can also influence agro-ecosystems environmental aspects; for this reason, the intensification of agricultural practices as required in the modern agriculture, especially related to soil tillage and fertilization practices, represent a serious issue that should be adequately approached to avoid losses of soil fertility and quality associated with elevated greenhouse gas emissions, mainly carbon dioxide [? ]. According to this perspective, recently, Radicetti et al. [? ] underlined the importance of some agronomical techniques based on ecological approaches, such as cover cropping and conservation tillage, to enhance the soil C sequestration and to reduce soil CO 2 emissions. Although it is well known that the benefit of reduced tillage practices and fertilization source is their being environmentally friendly practices [? ? ? ? ], this is one of the few studies that focused on the effect of the soil tillage and fertilization source on soil CO 2 emissions and C balance during potato cultivation in the Mediterranean environment of central Italy. In fact, on one hand, the adoption of reduced tillage systems are strongly encouraged because they save money and labor, improve soil biological properties, and prevent soil compaction [? ] on the other hand, their effect on crop yield should be adequately evaluated to avoid severe yield losses.
The data showed that the marketable potato yield was significantly affected by soil tillage in both growing seasons, with the greater values observed in plowing and spading soil tillage compared to subsoiling. According to Mundy et al. [? ] subsoiling tends to increase soil bulk densities compared to conventional tillage, and it can decrease potato yield due to reduced potato root growth to mechanical resistance of the soil, as well as the reduced resistance to periods of water stress. Although subsoiling is encouraged under sustainable management of agro-ecosystems for improving soil biological and physical conditions that could potentially benefit soil quality, this practice seems not recommended for potato cultivation under soil conditions similar to the experimental site. The lowest marketable potato yield observed under subsoiling tillage associated with the highest cumulated CO 2 emissions, and C output determined also low C input/C output ratio and high CO 2 emission/yield ratio, making it doubtful as an environmental-friendly strategy for potato cultivation. However, Carter et al. [? ] reported similar potato yield among different soil tillage; probably the soil conditions of the experimental area do not allow the adoption of subsoiling for potato cultivation, even if further research should be carried out to evaluate the effect in other kinds of soil in the same environmental conditions. In any case, the C input/C output ratio (>1) among all tillage treatments indicates that plowing, subsoiling, and spading could serve as a temporary sink of C, in according with the findings of Mancinelli et al. [? ]. On the contrary, plowing and spading tillage showed high and similar potato yield in both growing seasons, even if under spading tillage conditions the C output resulted the lowest compared with the other tillage treatments. According to Giordano et al. [? ] in Mediterranean areas, the spading machine has a discrete success due to the absence of a compact layer at the bottom of the tilled soil and the reduced drawbar pull required. The results are in agreement with the findings reported by Laudicina et al. [? ], where the spading machine is effective in preserving soil organic matter pool and in improving soil quality; indeed spading treatment showed the highest C input/C output ratio and the lowest cumulated CO 2 emission/potato yield ratio. Furthermore, in suitable soil conditions, a single passage of spading machine could be enough for seed bed preparation reducing the fuel requirements and energy consumption as normally required with plowing. According to Usaborisut and Prasertkan [? ], field operations represent the main expensive operation in crop production and the adoption of inadequate soil tillage operations may result in an excess of fuel consumption and inefficient use of energy. Recently, Fox et al. [? ] showed that 7.5 cm of deeper tillage increased the fuel consumption of 50% without any increase in the crop yield.
The biggest yield of marketable potato tubers observed in 2016 could be due to well distributed precipitation throughout potato growing season and reduced temperature compared with 2015. In fact, potato is a shallow rooted crop extremely sensitive to water stress and high temperature, especially in late season when the potato canopy is dry and the soil is highly exposed to solar radiation [? ]. Increased emissions of CO 2 in the potato growing season were observed regardless to the applied experimental treatments; the increase could be related to the higher soil temperature and moisture during the potato cultivation, especially in 2016. In fact, the results showed that soil CO 2 emission rates were significantly correlated with volumetric soil water content and soil temperature and varied according to soil tillage and fertilizer source treatments.
The obtained results suggest that the incorporation of organic fertilizer increases soil CO 2 emissions compared with mineral fertilizers throughout the potato growing seasons. In agreement with Mancinelli et al.
[? ] the difference on soil CO 2 emission among the fertilizer sources could be due to the organic materials that affect biological properties and organic mineralization of the soil. Despite the high soil CO 2 emission rate measured under organic fertilizers, the high C input by biomasses ensure a high C input/C output ratio (>1). These results prove that by applying organic fertilizer, the soil could become a carbon sink, and this therefore represents a possible solution for reducing the soil CO 2 emissions in sustainable agro-ecosystems. Moreover, the use of organic urban waste as source of organic fertilizer is strongly encouraged because it increases the economic value of a waste material, and it is seen as a way of food production by adopting the principles of circular economy [? ]. In addition, the low cumulated CO 2 emission/yield ratio indicates that the same amount of potato tubers is produced with low soil CO 2 emissions in the mineral fertilization treatment. The tuber yield differences observed in 2016 could be due to the nutrient availability, mainly nitrogen, between the fertilization sources (data not shown). These results are in line with the findings of Campiglia et al. [? ], which observed similar yield between organic and mineral fertilization during the shortage period of rainfall and soil water content and higher yield in mineral than organic fertilization under high level of precipitation and soil water content. In fact, the nutrient availability under organic fertilization has high sensitivity on the mineralization process, driven also by weather conditions. On the contrary, mineral fertilizer was applied when the weather conditions were suitable for potato crop in order to supply nutrients at the right moment for the crops need [? ].

Conclusions
This study provides useful information on the behavior of soil tillage and fertilizer source on soil CO 2 emissions and C balance in potato cultivation agro-ecosystem under the Mediterranean environment in central Italy. In particular, this study highlights how it is possible to reduce soil CO 2 emissions in the agro-ecosystems and improve the sustainability of intensive systems. The findings showed that all soil tillage evaluated are environmentally friendly (input/output ratio >1), even if the potato yield results suggest to prefer spading tillage for seed bed preparation of potato crop. Indeed, through soil spading for potato seed bed preparation, it is possible to obtain adequate tuber yield and reduced carbon losses by carbon dioxide. Moreover, the results suggest that spading tillage and organic fertilization could achieve environmental benefits, in terms of reduced soil CO 2 emissions, and enhance agronomical benefits, in terms of marketable tuber yield of potato. Moreover, the findings are of practical application for the farmers because of reduced costs due to tillage by saving energy and fuel and valorizing organic waste in agricultural application. Further studies should be performed to evaluate also the effect on microbial biomass and soil carbon sink, with the aim to choose the most suitable agronomical practice for potato cultivation that could maintain crop yield and shift the carbon balance in favor of soil C accumulation over a long period. Moreover, these findings may also be extended to other environments and crops which have similar requirements to potato in terms of the crop's nutritional and environmental conditions.