¹⁵N Natural Abundance, Nitrogen and Carbon Pools in Soil-Sorghum System Amended with Natural and NH₄⁺-Enriched Zeolitites

Abstract

The use of rocks containing high amounts of natural zeolites (zeolitites) as soil amendment has been found as a valuable method for increasing agriculture sustainability. However, the potentialities and the effects of zeolitites on the biogeochemical cycles of nitrogen (N) and carbon (C) have still not been clearly addressed in the literature. The objective of this study was therefore to investigate the N and C pools and ¹⁵N distribution in an agricultural soil amended with both natural and NH₄⁺-enriched zeolitites with the aim of understanding their effects on the soil-plant system, during sorghum cultivation, under fertilization reductions. Zeolitites were applied to an agricultural soil both at natural state (5 and 15 kg m⁻²) and in an enriched state with NH₄⁺ ions from pig slurry (7 kg m⁻²). Both zeolitites at natural and enriched state increased soil cation exchange capacity and affected microbial biomass, causing an initial decrease of microbial C and N and then a possible increase of fungal population. N-NO₃⁻ content was lower in natural zeolitite treatments, that lead to a lower NO₃⁻ availability for denitrifying bacteria. Zeolitites slightly affected the fixed N-NH₄⁺ pool. δ¹⁵N turnover indicated that N from NH₄⁺-enriched zeolitites remained in the soil until the growing season and that fertilizers partially substituted the fixed pool. Leaf δ¹⁵N content indicated that plants assimilated N from NH₄⁺-enriched zeolitites and evidenced a higher fertilization recovery in natural zeolitite treatments. Organic C tended to be higher in all zeolitite treatment rhizospheres. In soils amended with zeolitites at natural state (at both application rates) sorghum yield was similar (+3.7%) to that obtained in the control while it was higher (+13.9%) in the plot amended with NH₄⁺-enriched zeolitites.

1. Introduction

The intensification of agricultural and zootechnical technology has brought food production to a high cost in terms of environmental quality [1,2,3,4]. The large use of fertilizers, as well as the need to apply manure, increase nutrients leaching and runoff risk to ground and surface water (e.g., eutrophication), soil desertification, biodiversity reduction and to an increase in greenhouse gases production (e.g., NH₃, N₂O, CO₂) [4,5].

In this respect, natural zeolitites (NZ), particularly those containing clinoptilolite and chabazite, have been widely used in agriculture and horticulture. Zeolitites are rocks containing more than 50% of zeolites [6], hydrated minerals capable of binding NH₄⁺ from solutions in their extra-framework sites, preventing NH₄⁺ leaching by rainfalls and/or irrigation. In this way, zeolitites can increase nitrogen use efficiency (NUE) and behave as slow-releasing fertilizers, resulting in a more sustainable agriculture, where the amount of fertilizers can be decreased without lowering the production yield [6,7,8,9,10,11,12]. Due to their physic-chemical characteristics, zeolitites can be used as soil conditioners, carriers of fertilizers, antibacterial agents, insecticides, pesticides, as enhancer of soil biological activity and additional reducer of the Sodium Adsorption Ratio (SAR) index [13,14,15,16,17,18,19]. A yield increases of up to 65% in several crops (sorghum, wheat, corn, rice, marrow, chard, basil, tomato, radish, watermelon, lettuce, sunflower) in soils amended with natural or NH₄⁺-enriched zeolitite (NEZ) were reported by several studies [6,20,21]. Analogously, Reháková et al. [14] found an increase of roots biomass in parsley (Petroselinum satilvum), carrots (Daucus carota), and onions (Allium cepa). The same authors detected an increase in strawberry (Senga sengana) yield and green biomass of 70% and 49%; respectively, when natural zeolitite was applied as fertilizer. The aforementioned studies provide very important and useful information regarding the beneficial effects obtained using NZ in terms of crop yield, environmental protection (reduction of nitrate leaching) and water use efficiency. Notwithstanding these promising results, a detailed study on the N and C pools dynamics in a zeolitite-amended field is still lacking.

The implementation of the Water Framework Directive (2000/60/EC) and the Nitrate Directive (91/976/EEC) led to the designation of large areas vulnerable to nitrate pollution (Nitrate Vulnerable Zone, NVZ) such as the Ferrara Province (Emilia Romagna, Italy). It lays at the end of the Po Plain where the intensive agricultural practices of the entire Emilia-Romagna are conveyed. The high number of breeding farms located on this territory and the consolidate agricultural practice of animal slurry spreading (liquid mixture of urine, water and feces), increases N₂O emission by nitrification/denitrification and NO₃⁻ leaching, further emphasizing the vulnerability of this province [22].

In this framework, ZeoLIFE project (LIFE10+ ENV/IT/00321) was conceived to test different zeolitite treatments (ZT) at the field scale aiming at reducing the input of N from fertilizers and the irrigation water in agriculture. This project exploits the high cation exchange capacity (CEC) of this material by up-taking NH₄⁺ from pig slurry using an especially designed prototype tank [23] or, alternatively, from chemical fertilizers directly in the field after addition to the soil [24].

The current scientific literature lacks field studies in which a detailed investigation of the dynamics of N and C pools in soil modified with natural zeolitites was conducted. In this paper, we present a detailed study on the dynamics, of both organic and inorganic N and C pools in soils amended with NZ and NEZ (e.g., exchangeable and fixed NH₄⁺ in clay interlayers, nitrate, microbial biomass N and C, total organic carbon and carbon in humic substances).

Conventional agricultural practices (plowing, application of chemical fertilizers and monocropping) were compared with those related to the innovative zeolitite methods. Non-amended soils (control) were compared with both NZ and NEZ amended soils (applied at different rates), over a cultivation cycle of sorghum (Sorghum vulgare Pers.). The main objectives of this work were therefore the following:

(1)

Evaluate how the addition of NZ and NEZ affects N and C pool dynamics in the soil-plant system over one agronomic year;

(2)

Determine if the use of NZ or NEZ can allow a reduction in the application of chemical fertilizers to soil;

(3)

Determine if the use of NZ or NEZ have beneficial effects on crop yield.

Considering the global need of improving agricultural sustainability and reducing environmental pollution, this work may help to unveil the potential of these methodologies.

2. Materials and Methods

2.1. Zeolitites

The natural K-rich, Na-poor zeolitite used in this study is a byproduct from a quarry located near Sorano village (Italy, Grosseto province) (Figure 1, 42°41′20.65″ N; 11°44′26.29″ E) that is mainly exploited to obtain blocks and bricks for construction and gardening. It is a low-cost, natural and eco-friendly granular by-product of the quarrying activity, part of a large zeolitized pyroclastic deposit whose total zeolitic content is on average 70% (chabazite, 68.5%; Phillipsite, 1.8%; Analcime, 0.6%; [10]). The total CEC of the zeolitite is 2170 mmol kg⁻¹, of which 1460 are due to Ca²⁺, 600 to K⁺, 70 to Na⁺ and 40 to Mg²⁺. The dry bulk density and water retention vary with grain size from 870 kg m⁻³ and 48.4% for Ø <3 mm to 560 kg m⁻³ and 34.2% for Ø 3–6 mm, respectively.

The grain size distribution of the employed zeolitites are reported in Figure 2. Part of the NZ was subjected to an enrichment process, which allowed the saturation of the zeolitite with NH₄⁺, creating the NEZ. The enrichment process was carried out by static mode in a prototype tank where the zeolitite is mechanically mixed with pig slurry (average NH₄⁺ content of 2000 mg L⁻¹) at specific solid/liquid ratio, stirring and resting times, gaining between 400 and 800 mmol of NH₄⁺ per kg of material [23]. N and C pools of this material are reported in

Table 1. Average N and C pools of NH₄⁺-enriched zeolitite (NEZ). Standard deviation within brackets.

TN	g·kg−1	5.84 (0.90)
Fix N-NH4+	mg·kg−1	343 (21)
Exch N-NH4+	mg·kg−1	4884 (330)
N-NO3−	mg·kg−1	148 (7.4)
Nmic	mg·kg−1	876 (152)
TOC	g·kg−1	10.3 (0.1)
Cmic	mg·kg−1	117 (1)
C/N		1.78
Cmic/Nmic		0.13

.

2.2. Experimental Field Setting

The experimental field is located near Codigoro town (Italy), 40 km eastward of Ferrara (Figure 1, 44°50’33’’ N and 12°05’45’’ E) and 15 km from the Adriatic Sea in a reclaimed area at an average altitude of 3 m below sea level. The average daytime temperatures ranges from 3°C in January and 25 °C in July and the marine thermoregulation generally maintains the minima over zero, reducing the number of night frosts. The average rainfall is between 500 and 700 mm per year, representing the regional pluviometric minimum with peaks in autumn and summer (sub-continental climate).

The area belongs to the eastern Po River plain, where ancient interdistributary bays and brackish marshes were recently (1860–1960) reclaimed [25,26]. The soil of the experimental field belongs to the cartographic unit FOR1-LCO1 of the Emilia Romagna 1:50,000 Soil Map (www.regione.emiliaromagna.it), and consisting of recent interfluvial silty-clay deposits classifiable as Calcaric Gleyic Cambisol [26,27]. The average grain size of the upper soil horizon is constituted by 0.6 ± 0.1% of medium sand (200–630 µm), 7.4 ± 0.3% of fine sand (63–200 µm), 49.2 ± 3.1 of silt (2–63 µm) and 42.0 ± 3.4% of clay (<2 µm) [28]. The average organic matter content is 8.1 ± 1.5%, hydraulic conductivity of the soil is 1.7 ± 2.4 m day⁻¹ while the average bulk density is 1.15 ± 0.05 kg m⁻³ [28]. The upper soil horizon is also generally well drained due to the presence of a sub-surface drainage system at −1 m.b.g.l. Soil mineralogical composition is characterized by quartz, feldspar, calcite and clay minerals (illite, smectite, clorite, serpentine and mixed-layer) [10].

The experimental field was parceled in order to compare the different zeolitite treatments with the traditional practices. The parcels were designed linear and continuous in order to facilitate the movements of farm machines. In each parcel, three different sub-parcels were considered to consider inter parcel variability and to have statistical significance. One parcel (0.5 ha) was amended with 7 kg m⁻² (7CZ) of NEZ (Ø < 3mm); two parcels (1 ha each) were amended with 5 (5NZ) and 15 (15NZ) kg m⁻² of coarse-grained (Ø 3–6mm) NZ, respectively, and 3.5 ha were used as control (CNTR). The smaller size of 7CZ parcel was due to the long time required to produce the NEZ using the “ZeoLIFE” prototype tank [23,24], which has a limited production rate of about 250 kg day⁻¹. Addition and spreading into the field of NZ and NEZ were carried out between October 7th and November 6th, 2012. NEZ supplied approximately 410 kg N ha⁻¹ in the 7CZ parcel. Immediately after spreading, the field was ploughed and the zeolitite was located until an approximate depth of 30–40 cm. According to ZeoLIFE plan, in the first year of field experiment, sorghum (Sorghum vulgare Pers.) was sowed on May, 9^th, 2013. Nitrogen fertilizers were distributed in two steps: di-ammonium phosphate during the sowing, with an application of 21.6 kg N ha⁻¹ in all parcels and urea on June 3^rd, 2013, with different doses in the various parcels. In the CNTR parcel, an average of 170 (±16) kg N ha⁻¹ were supplied, while a reduction of approximately 30% and 50% of urea in NZ (both 5NZ and 15NZ for 122 and 127 kg N ha⁻¹, respectively) and NEZ parcels (81 kg N ha⁻¹) was applied, respectively. Harvest occurred in September, 23^rd, and the yield was separately evaluated for each treatment.

2.3. Soil and Plant Sampling

A total of 60 soil samples (from 0 to 30 cm depth) were collected in the three sub-parcels for each treatment by manual drilling using an Ejielkamp Agrisearch auger. Soil samplings were carried out in four periods of the agronomical year: (1) November 2012 (Nov-2012), after the zeolitite application; (2) May 2013 (May-2013), before sowing and the first fertilization; (3) June 2013 (Jun-2013), after the urea addition; (4) September 2013 (Sep-2013), at the harvest. Soil samples collected at the harvest were subdivided in bulk soil and rhizosphere. The rhizosphere was separated by gently shaking the plant roots. Three representative plants from each treatment and each sub-parcel were sampled during the harvest; each plant was subdivided in roots, stems, leaves and panicles that were subsequently dried at 50 °C and milled (total of 48 samples).

2.4. Analytical Techniques

CEC and exchangeable bases were determined using the Co(NH₃)₆Cl₃ method [29,30] on both bulk soil and rhizosphere samples collected in Sep-2013, pH was determined on 1:5 (w:v) extracts using a pH-meter, while carbonate content was determined using the volumetric method. Total nitrogen (TN) and organic C (TOC) of soil and plant samples were determined by an elemental CHNS-O EA 1110 Thermo Fisher Scientific analyzer coupled with mass spectrometry (Delta Plus, Finnegan, Thermo-Fisher). Soil samples were weighted in silver pots and treated with HCl to eliminate the inorganic C, while plant samples were weighted in tin pots. ¹⁵N natural abundance was measured in total and fixed nitrogen pools (TN and Fix N-NH₄⁺) and in sorghum organs; the natural abundance was expressed in δ‰ (Equation (1); [31]), where δ represents the difference from the standard atmospheric N₂ (0.3663% ¹⁵N):

$$\mathsf{\delta }{}_{}{}^{15}\mathrm{N}(‰)=\left[\frac{\left(\frac{{}_{}{}^{15}N}{{}_{}{}^{14}N}\right)sample- \left(\frac{{}_{}{}^{15}N}{{}_{}{}^{14}N}\right)st}{\left(\frac{{}_{}{}^{15}N}{{}_{}{}^{14}N}\right)st}\right]\times 1000$$

(1)

To evaluate the influence of the N source on the plant N isotopic signature, mixing lines were built between each end-member represented by soil, NEZ and urea.

NO₃⁻ was extracted with Milli-Q (Millipore USA) water in a 1:5 (w/v) ratio, the solution was shaken for 1 hour and then filtered. NO₃⁻ was determined by ion chromatography using an isocratic dual pump ion chromatography ICS-1000 Dionex. An AS-40 Dionex auto sampler was employed to run the analysis; Quality Control (QC) samples were run every 10 samples and the standard deviation for all QC samples was better than 4%.

Exchangeable N-NH₄⁺ was extracted with 1M KCl in a 1:10 (w/v) ratio, the solution was shaken for 1 hour and then filtered. The solution was diluted and analyzed with an Ion Selective Electrode (ISE) Orion 95-12 connected to a Thermo Fisher Orion 4star pH-ISE benchtop. Interferences with Cl⁻ and K⁺ were verified and excluded by comparing measurement made with NH₄Cl, CH₃COONH₄ + 1M KCl and CH₃COONH₄ + H₂O standard solutions.

After the elimination of organic and exchangeable N, fixed N-NH₄⁺ (Fix N-NH₄⁺) was determined by adding 20 mL of KBrO to 2 g of air-dried soil samples in two replicates. After 2 hours, 40 mL of distilled water was added and a heating cycle of 10 min in a microwave oven was applied. Then, 20 mL of Milli-Q H₂O was added and samples were left still for 24 hours. Finally, samples were centrifuged at 5000 rpm for 10 min and washed 2 times with 0.5 M KCl and analyzed by elemental analysis for the evaluation of TN and δ¹⁵N [32].

Total extractable carbon (TEC) was obtained adding 100 mL of 0.1 M NaOH + 0.1 M NaP₂O₇·10H₂O to 10 g of air-dried soil. The solution was left in a Dubnoff bath at 60°C for 24 hours and then centrifuged at 6600 rpm for 15 min and filtered at 0.45 mm (Millipore). Part of this solution was acidified with HCl 1:1 to reach pH < 2 to allow the precipitation of Humic Acids (HA), which were separated after centrifugation at 7000 rpm for 25 min by the Non-Humic-Substances (NHS) and Fulvic Acids (FA). The separation of FA from NHS was performed by solid chromatography using a polyvinylpyrrolidone (PVP) polymer in acid ambient, which allows the retention of FA on the PVP and lets NHS pass through the column. The FA was then recovered by bringing the pH of the column to alkaline values with 0.5 M NaOH. The evaluation of C contained in TEC, HA and FA was obtained by reverse titration with 0.2 N Fe₂SO₄ after oxidation of the C contained in the sample to CO₂ adding 5 mL of 1 N K₂Cr₂O₇ + 20 mL of H₂SO₄ and heating at 150 °C for 10 min [33]. The C extraction was carried out on one replicate.

Humification Index (HI) and Humification Degree (HD) were calculated as the ratio of C contained into NHS divided by the C contained into Humic Substances (HS) and as the ratio between HS and TEC expressed as percentage respectively (Equations (2) and (3), [33,34]).

$$HI=\frac{\left[TEC-\left(HA+FA\right)\right]}{HA+FA}$$

(2)

$$HD=\frac{\left[100\times \left(HA+FA\right)\right]}{TEC}$$

(3)

where HI is the Humification Index, DH is the Humification Degree, TEC is Total Extractable C, HA is Humic Acids and FA are Fulvic Acids.

C and N immobilized by microbial biomass (C_mic and N_mic) were obtained by the chloroform fumigation-extraction method [35]. Following this, 10 g of soil samples in four replicates were humidified until 60% of water holding capacity (WHC) and incubated at room temperature for 1 week at constant WHC. After this period, two replicates were extracted with 0.5 M K₂SO₄, representing the non-fumigate sample (NF). On the other two samples, CHCl₃ was added in a closed drier equipped with a void pump to allow CHCl₃ volatilization at room temperature. In these fumigate (F) samples the presence of CHCl₃ atmosphere guarantees the death of the microbial biomass and allows release of the immobilized C and N after the same extraction with 0.5 M K₂SO₄ performed for NF. After the extraction, both NF and F samples were analyzed with a Shimadzu Total Organic Carbon Analyser TOC-V CPN coupled with a TN unit (TNM-1). Sorghum harvest was carried out with a harvester and the total yield of each parcel was weighted separately directly in the farm using the local vehicle scale. It was not possible to harvest separately the three sub parcels of each treatment, as this practice would have required an excessive time and precision not compatible with the host farm agricultural practices.

2.5. Statistical Analysis

At least three replicates per treatments have been analyzed at each sampling point to have statistical significance. For evaluating significant differences between the treatments, parametric statistic was applied to the dataset. Two-Way ANOVA was employed (after verifying normality of data distribution with Shapiro-Wilk test) for testing significant differences within factors time and treatments and for testing interactions between factors (treatments*time). Successively, a series of One-Way ANOVA and Tukey (HSD) post-hoc pairwise multiple comparison tests were applied at each sapling time at “p” level of 0.05 using Sigmaplot 12.0. Correlation analyses were also performed (using Pearson coefficient “ρ”) in order to quantify linear associations between variables for each treatment.

3. Results

After the enrichment process with pig slurry, NEZ gains a total N load of 5.84 (± 0.90) g kg⁻¹ (Table 1). Most of the N was present in the form of Exch N-NH₄⁺, while Fix N-NH₄⁺ is considerably lower. N-NO₃⁻ load is on average 148 mg kg⁻¹ (the material is not washed after the treatment before the introduction into the soil system). NEZ is characterized by a very high N_mic and a low C_mic, resulting in a C_mic/N_mic ratio lower than 1. TOC content is quite low (10.3 g kg⁻¹ on average) resulting in a low C/N. Soil pH was not affected by the addition of both NZ and NEZ (p > 0.05), remaining close to neutral values in all treatments (with exception of Jun-2013 where after urea addition an increase toward sub-alkaline values (p < 0.05), Figure 3). CEC increased significantly (p < 0.05) in the bulk soil of all ZT. Ca²⁺ and Mg²⁺ were the major exchangeable bases, followed by K⁺ and Na⁺, with the latter always higher in 7CZ than in the other treatments (

Table 2. Average of exchangeable bases and cation exchange capacity (CEC) of soil samples, both bulk soil and rhizosphere. Standard deviation within brackets. Significant (p < 0.05) or non-significant (p > 0.05) differences between treatments at each sampling time are indicated by capital letters.

		Ca2+	K+	Mg2+	Na+	CEC
Time	Treatment	cmol·kg−1	cmol·kg−1	cmol·kg−1	cmol·kg−1	cmol·kg−1
Sep-2013 Rhizo	CNTR	18.3 (0.7) A	8.3 (1.7) A	14.4 (1.2) A	1.8 (0.7) AB	42.8 (4.0) A
	5NZ	18.6 (0.6) A	9.9 (0.8) A	17.2 (0.9) B	1.3 (0.1) A	47.1 (1.9) A
	15NZ	17.7 (0.9) A	10.4 (0.8) A	17.9 (1.4) B	1.6 (0.6) A	47.5 (2.5) A
	7CZ	19.4 (1.1) A	7.7 (3.8) A	15.9 (0.6) AB	3.0 (0.4) B	46.0 (2.7) A
Sep-2013 Bulk	CNTR	19.9 (1.3) A	6.1 (3.3) A	16.8 (1.3) A	2.5 (1.1) AB	45.3 (2.3) A
	5NZ	20.3 (0.01) AB	4.9 (0.1) A	21.9 (0.1) B	2.0 (0.02) A	49.1 (0.04) AB
	15NZ	23.0 (0.02) B	5.9 (0.004) A	19.5 (0.1) AB	1.7 (0.1) A	50.2 (0.2) B
	7CZ	22.0 (0.2) AB	5.0 (0.1) A	19.9 (0.2) B	3.6 (0.04) B	50.6 (0.6) B

). No significant differences were found between the various treatments regarding exchangeable K⁺ (p > 0.05), while Ca²⁺ and Mg²⁺ are generally higher in ZT, especially in the bulk soil with respect to the rhizosphere (p < 0.05).

3.1. Nitrogen and Carbon Pools in Soil

Soil N pools significantly differed (p < 0.05) between the treatments and among the sampling times, with the exception of TN that shows significant differences among sampling times only (p < 0.05) (Table S1, Supplementary Materials). High TN were detected in Nov-2012 and Sep-2013, while a depletion was observed in May-2013 and Jun-2013.

Similar behavior is recorded for Fix N-NH₄⁺ pool but in this case, all ZT shows slightly lower values (p < 0.05) in Nov-2012 with respect to the CNTR, while in Jun-2013 only 15NZ treatment is significantly lower (Table S1, Supplementary Materials). At the beginning of the experiment, N-NO₃⁻ content was significantly higher in the CNTR and 7CZ parcels than in those amended with NZ (5NZ and 15NZ, respectively) (p < 0.05). In May-2013 a general depletion of N-NO₃⁻ pool is noticed in all the treatments, increasing in Jun-2013 after urea addition especially in CNTR parcel. N-NO₃⁻ values of ZT (especially 5NZ and 15NZ) were remarkably lower in this period with respect to the CNTR (p < 0.05) (Table S1, Supplementary Materials). Exch N-NH₄⁺ was significantly affected by NEZ addition in Nov-2012 (p < 0.05) (138 mg kg⁻¹ vs 50 mg kg⁻¹ of 7CZ and CNTR, respectively) while starting from May-2013 these differences completely disappeared (p > 0.05). In Jun-2013, there were no differences between the treatments (p > 0.05) while a relative enrichment of the bulk soil Exch N-NH₄⁺ with respect to the rhizosphere was apparent in Sep-2013.

Generally, at the beginning of the experiment, all treatments showed a low N_mic content, following the order 7CZ > CNTR > 5NZ = 15NZ, suggesting that zeolitites differentially affected the N_mic pool. N_mic was increased by NEZ addition while it was decreased by NZ addition. The N_mic pool increases in May, while it decreases in June after the chemical fertilization; as expected, N_mic content in rhizospheric soil is higher than that determined in the bulk soil at the harvest (Table S1, Supplementary Materials). A remarkable increase of C_mic and N_mic and a contemporaneous decrease of N-NO₃⁻ in all the treatments was observed in May-2013 (Figure 4).

During the monitored period, C_mic ranged from 16 to 432 mg kg⁻¹. At the beginning of the experiment large differences (p < 0.05) were detected among the various treatments. In this period, C_mic content decreases following the order CNTR (148 mg kg⁻¹) >15NZ (98.5 mg kg⁻¹) > 7CZ (43.2 mg kg⁻¹) > 5NZ (16.1 mg kg⁻¹). In May-2013, C_mic increased in all treatments, except for 15NZ, and decreased again in Jun-2013. As expected, C_mic content in the rhizosphere is higher than that determined in the bulk soil. The average C_mic/N_mic ratio ranges from 3.3 to 94 (Figure 5). In Nov-2012, low values of C_mic/N_mic were observed in 5NZ and 7CZ treatments; concerning 5NZ, the low ratio is due to the lower C_mic with respect to the other treatments, while for 7CZ the low ratio may have been due to the high N_mic. All ZT have a higher C_mic/N_mic ratio than CNTR after May-2013 because of the lower N_mic; on average, C_mic/N_mic ratio increased in order of 15NZ > 5NZ > 7CZ > CNTR. TOC showed significant differences between the treatments and among the sampling times (p < 0.05). In particular, a depletion of TOC content was observed in 7CZ in Nov-2012 and in Jun-2013, after urea addition (Table S1, Supplementary Materials). As expected, TOC content in Sep-2013 was higher in rhizosphere than in the bulk soil and there was a significant correlation between TN and TOC (ρ = 0.83, p < 0.05). Soil C/N ratio showed significant differences among the treatments only at the beginning of the experiment because of the higher TN of 7CZ (p < 0.05), while during the rest of the monitored period, soil C/N ratio remained constant around 10 (Table S1, Supplementary Materials). HS C content was reported in

Table 3. C content of Humic Acids (HA) and Fulvic Acids (FA), percentage of Total Extractable and Total Organic Carbon (TEC/TOC) ratio, Humification Index (HI) and Humification Degree (HD) according to Ciavatta et al. [33] and Sequi et al. [34].

		HA	FA	TEC/TOC	HI	HD
Time	Treatment	g kg−1	g kg−1	%		%
Nov-2012	CNTR	5.3	3.4	47.4	0.4	34.4
	5NZ	4.3	5.0	46.2	0.2	37.7
	15NZ	3.9	3.7	41.4	0.5	27.8
	7CZ	2.7	3.9	40.1	0.3	31.9
May-2013	CNTR	7.4	3.2	63.0	0.5	43.1
	5NZ	8.6	3.5	74.4	0.6	46.9
	15NZ	11.4	3.6	87.3	0.3	68.2
	7CZ	5.9	2.5	48.6	0.1	43.8
Jun-2013	CNTR	4.5	3.1	52.3	0.4	38.6
	5NZ	4.3	2.9	65.0	0.8	35.5
	15NZ	3.8	3.1	50.6	0.3	39.2
	7CZ	3.4	3.0	43.6	0.6	27.4
Sep-2013 Bulk	CNTR	5.9	3.2	52.9	0.2	44.2
	5NZ	3.9	2.9	54.5	0.9	29.4
	15NZ	8.6	3.3	70.7	0.4	51.3
	7CZ	5.9	3.5	57.9	0.4	41.2
Sep-2013 Rhizo	CNTR	8.1	3.1	58.1	0.3	43.4
	5NZ	6.9	3.2	48.5	0.4	34.7
	15NZ	9.7	3.2	49.2	0.2	41.0
	7CZ	7.1	3.3	48.8	0.3	37.0

. TEC content ranged from 8.3 to 19.2 g kg⁻¹ (data not shown), whereas the extraction ratio ranged between 41 to 87%. Notably, the highest extraction ratio was observed in 15NZ parcel, followed by 5NZ. Generally, HA C content is higher than that determined in FA, with exception of 7CZ in Nov-2012, in which the FA/HA ratio is higher than 1. A low HI was generally recorded in all the treatments with the exception of 5NZ, where especially in May, June and Spet-2013 Bulk, HI increases through values above 0.5, highlighting a prevalence of NHS. The HD% generally varied between 30% and 50%, with the exception of 15NZ that showed large variations (from 27.8% in Nov-2012 to 68.2% in May-2013).

3.2. Total N and C Content in Sorghum Plant Components

No significant differences in TN and TOC content were detected in the sorghum organs (e.g., roots, stems, leaves and panicles) among the treatments (p > 0.05) (Table S2, Supplementary Materials). Generally, leaves and panicles had a higher TN content than roots and stems, while TOC was similar in all the plant organs, resulting in decline of the C/N ratio following the order stems > roots > panicles > leaves.

3.3. δ¹⁵N Natural Abundance in Soil and Sorghum Organs

TN and Fix N-NH₄⁺ δ¹⁵N differs significantly (p < 0.05) among treatments and time (

Table 4. Average ¹⁵N natural abundance in soil N pools (TN and Fix N-NH₄⁺). Standard deviation within brackets. Significant (p < 0.05) or non-significant (p > 0.05) differences between treatments at each sampling time are indicated by capital letters. “*” and “x” symbols indicates the presence or the absence, respectively, of a significant difference among the variables for Treatments and Time in the experiment period and correlating treatments with time (Treatment*Time).

		TN δ15N vs air	Fix δ15N vs air
Time	Treatment	‰	‰
Nov-2012	CNTR	3.86 (0.15) A	6.00 (0.17) B
	5NZ	3.83 (0.22) A	5.28 (0.45) AB
	15NZ	3.38 (0.19) A	4.55 (0.31) A
	7CZ	7.46 (0.53) B	8.27 (0.36) C
May-2013	CNTR	3.75 (0.20) AB	3.34 (0.15) A
	5NZ	3.34 (0.06) A	3.38 (0.30) A
	15NZ	3.31 (0.62) A	3.84 (0.04) A
	7CZ	4.87 (0.17) B	3.11 (0.12) A
Jun-2013	CNTR	3.11 (0.28) A	3.65 (0.43) A
	5NZ	3.37 (0.10) A	3.38 (0.17) A
	15NZ	3.32 (0.05) A	3.46 (0.13) A
	7CZ	3.27 (0.17) A	3.73 (0.19) A
Sep-2013 Bulk	CNTR	2.68 (0.11) A	1.37 (0.11) A
	5NZ	2.39 (0.09) A	1.83 (0.23) AB
	15NZ	2.73 (0.14) A	1.79 (0.03) AB
	7CZ	2.55 (0.29) A	2.09 (0.16) B
Sep-2013 Rhizo	CNTR	4.22 (0.38) A	2.47 (0.26) A
	5NZ	4.17 (0.31) A	2.81 (0.75) A
	15NZ	4.55 (1.11) A	2.40 (0.46) A
	7CZ	4.45 (0.19) A	2.52 (0.31) A
NEZ		18.4 (0.5)	17.3 (0.4)
Urea		−1.18 (0.06)	-
Diammonium Phosphate		−0.89 (0.09)	-
	Treatment	*	*
	Time	*	*
	Treatment*Time	*	*

). At the beginning of the experiment (Nov-2012), 7CZ shows a remarkably higher isotopic signature (δ¹⁵N of 7.46‰ and 8.27‰ for TN and Fix N-NH₄⁺, respectively) with respect to the other treatments (p < 0.05). Furthermore, without considering 7CZ treatment, Fix N-NH₄⁺ pool showed a high δ¹⁵N variability (from 4.55‰ to 6.00‰) with respect to the TN pool (from 3.38‰ to 3.86‰). In May-2013, TN δ¹⁵N is still significantly higher (p < 0.05) in 7CZ treatment, while no significant differences were observed regarding Fix N-NH₄⁺ pool. In Jun-2013 (after the chemical fertilization), all treatments showed a homogeneous δ¹⁵N in both TN and Fix N-NH₄⁺ pools (p > 0.05). In Sep-2013, rhizosphere samples show higher δ¹⁵N with respect to the bulk soils in both TN and Fix N-NH₄⁺ pools (especially in the latter). The isotopic signatures of NEZ, urea and di-ammonium phosphate were: +18.4 (± 0.5) ‰, −1.18 (± 0.06) ‰ and −0.89 (± 0.09) ‰, respectively (

Table 5. Average ¹⁵N natural abundance in sorghum organs at the harvest. SD within brackets. Significant (p < 0.05) or non-significant (p > 0.05) differences between treatments for each organ are indicated by capital letters.

		δ15N vs Air
Plant Organ	Treatment	‰
Roots	CNTR	10.2 (1.4) AB
	5NZ	6.33 (1.69) A
	15NZ	10.3 (2.8) AB
	7CZ	11.6 (2.7) B
Stems	CNTR	9.95 (0.81) B
	5NZ	6.08 (1.56) A
	15NZ	11.1 (1.1) B
	7CZ	11.0 (1.5) B
Leaves	CNTR	12.3 (1.6) B
	5NZ	8.88 (0.94) A
	15NZ	13.8 (2.2) BC
	7CZ	18.1 (1.7) C
Panicles	CNTR	10.3 (0.4) B
	5NZ	6.82 (0.97) A
	15NZ	10.0 (1.8) B
	7CZ	11.7 (2.0) B

). Isotopic analysis on the harvested plants indicated that the δ¹⁵N of plant tissues was influenced by the different agricultural managements employed in the experimentation (Table 5). δ¹⁵N was significantly higher in 7CZ plants leaves and significantly lower in all 5NZ plants organs with respect to the CNTR plants (p < 0.05). In Figure 6, the TN vs δ¹⁵N of the different plants’ organs are plotted together with urea, NEZ and bulk soil. It is evident that the 7CZ leaves are strongly affected by NEZ contribution, while 5NZ plants are more influenced by a low δ¹⁵N N-source.

3.4. Sorghum Yield

The CNTR parcel returned a yield of 5818 kg ha⁻¹. The two NZ returned 6032 kg ha⁻¹ each, while 7CZ parcel gave 6627 kg ha⁻¹. Compared with the CNTR, the three zeolitite-treated parcels gained from 3.7 to 13.9%, although the yield results are not statistically significant.

4. Discussion

As expected, CEC was generally higher in ZT with respect to the CNTR, especially in the bulk soil. Ozbahce et al. [36] and Gholamhoseini et al. [20] found a similar tendency after mixing zeolitite in the first 30 cm of soil or after the spreading in combination with cattle manure. Both bulk and rhizosphere soils in the 7CZ parcel showed higher exchangeable Na⁺ with respect to the other treatments, likely because of the NEZ enrichment process with pig slurry.

4.1. NEZ Influence on Soil N and C Pools.

4.1.1. NEZ Effects on N Pool Size and Isotopic Signature

The addition of NEZ into the soil (7CZ treatment) did not lead to significant differences in TN or Fix N-NH₄⁺ content regardless the N input of about 410 kg of N ha⁻¹ brought by NEZ introduction. On the other hand, this N input has clearly influenced Exch N-NH₄⁺, N-NO₃⁻ and N_mic pools, increasing their size. Through the application of a simple mass balance equation, it was estimated that the increment in 7CZ Exch N-NH₄⁺ pool corresponded to the amount of N introduced with NEZ in the first 30 cm of soil.

As it is known, NH₄⁺-enriched zeolites can act as slow-released fertilizer [37,38,39] and in our experiment this behavior was clarified by the isotopic signature of the TN pool during the agricultural management. The δ¹⁵N increase of TN pool in 7CZ with respect to that found in the other treatments is in fact due to the N isotopic signature of the pig slurry contained into the NEZ. Dittert et al. [40] found a δ¹⁵N of 5.90‰ in agricultural soil treated with slurry while a δ¹⁵N of 3.90‰ was found in soils treated with inorganic fertilizers. Pig slurry usually has high δ¹⁵N values, as a result of the fractionation processes occurring during volatilization of NH₃ after the excretion, which causes an enrichment in the heavier ¹⁵N atoms in the substrate (slurry) with respect to the product (ammonia gas) [40,41,42,43,44]. The significantly higher δ¹⁵N of TN in May-2013 suggested that a significant amount of N introduced with NEZ was still present in the soil before the sowing. At the beginning of the experimentation, the organic C cycle of 7CZ treatment was affected by the high N input provided by NEZ, resulting in a decrease of TOC content and C/N ratio; furthermore, an increase of FA fraction can be interpreted as an increase in depolymerization processes.

4.1.2. NEZ Specific Effects on Soil Microbial Biomass

On the other hand, N_mic increment in 7CZ parcel was not counterbalanced by a parallel C_mic increment, resulting in a strong decrease of the C_mic/N_mic and likely in an increase of organic matter (OM) mineralization processes [44]. No evidence of enhanced nitrification was observed in 7CZ parcel after the zeolitite spreading as N-NO₃⁻ content was very similar to the CNTR notwithstanding the large N input provided by NEZ spreading (with residual N-NO₃⁻ from the pig slurry treatment). The agronomic year before the experimental cultivation was low in precipitation and the drought caused low yield and accumulation of high amounts of N from fertilizers in the upper soil horizon, thus no differences were expected between CNTR and ZT. A possible explanation of this behavior is that controlled retention and release capacity of NH₄⁺ contained in the microsite of the open-ring structure of the mineral. These ions are physically protected and likely less available to nitrifying bacteria (leading consequently to a lesser N-NO₃⁻ content in the pore-waters) [45,46,47,48]. This hypothesis is plausible but not supported by a direct measurement of gross nitrification rates (e.g., by the mean of ¹⁵N pool dilution technique). In this light, further studies are urgently needed where gross N transformation rates are measured in soil amended with NEZ and NZ to efficiently quantify gross N production and consumption through mineralization, nitrification and denitrification processes.

In May-2013, C_mic and N_mic increased significantly in all treatments, while N-NO₃⁻ pool significantly decreased (but not Exch N-NH₄⁺ pool with the exception of 7CZ) (Figure 4), suggesting that microbial immobilization prevailed over mineralization [49]. In addition to NO₃⁻ decrease, 7CZ treatment showed a reduction of Exch N-NH₄⁺ pool, where the N surplus caused by the NEZ introduction completely disappeared. A fraction of this mineral N depletion can be attributable to NO₃⁻ leaching and/or to N₂-N₂O losses through denitrification process (probably enhanced by waterlogged conditions; [25]). It is important to consider that NEZ introduction affected mainly the Exch N-NH₄⁺ and not the Fix N-NH₄⁺ pool and that 7CZ still had high TN δ¹⁵N in May-2013, confirming that an important fraction of NEZ-N was still present in the soil system. It is thus possible that a large fraction of the exchangeable N contained into NEZ has been immobilized by the microbial biomass [50,51].

4.1.3. NEZ Specific Effects on Fixed N-NH₄⁺ Pool

A possible hypothesis is that Fix N-NH₄⁺ pool is in dynamic equilibrium with exchangeable and soluble N-NH₄⁺ pools [52,53]. In our case, this hypothesis is supported by the δ¹⁵N turnover of fixed pool that suggested a continuous N replacement during the monitoring period. At the start of the experimentation, Fix δ¹⁵N of CNTR and NZ treatments was probably still affected by previous slurry treatments as shown by the higher isotopic signature. Generally, microbial activity increases with the increase of soil temperature, affecting the other N pools, in particular, a decrease in N-NO₃⁻ and Exch N-NH₄⁺ pools (as in May-2013 sampling) can alter the equilibrium with fixed interlayer ammonium. This decrease can cause a release of N from Fix N-NH₄⁺ pool in order to maintain equilibrium, with a continuous N turnover and a consequent change in δ¹⁵N. No δ¹⁵N turnover of fixed ammonium was observed immediately after the chemical fertilization applied in Jun-2013, while the addition of urea was marked by a pH increase [54]. On the contrary, in Sep-2013, a probable storage of N from chemical fertilization was visible in Fix N-NH₄⁺ pool (and consequently in TN pool), as demonstrated by the significantly lower δ¹⁵N (very close to the δ¹⁵N of urea and di-ammonium phosphate). This behavior is recorded only in the bulk soil, where the exploitation by the plant roots is less important than in the rhizosphere.

4.2. NZ Influence on Soil N and C Pools

The introduction of natural zeolitites (5NZ and 15NZ, respectively) affected the soil C and N pools. The decrease in the total reserve of Fix N-NH₄⁺ in 15NZ in Nov-2012 was probably due to a dilution effect operated by the high amount of NZ (devoid of N) introduced. Instead, the decrease of Fix N-NH₄⁺ during the growing season suggest an exploitation by plants, confirming the active role of this pool in crop nutrition [55,56]. Nevertheless, no significant differences in the amount of Fix N-NH₄⁺ pool were found between the various treatments, suggesting that the N subjected to CEC processes by zeolitites belongs to the exchangeable pool. The generally lower N-NO₃⁻ content in the NZ treatments may be attributed both to the aforementioned controlled retention of NH₄⁺ by the zeolitite, which can reduce the availability of N to nitrifying bacteria and to the lower amount of N fertilizer (from 30 to 50%) with respect to CNTR (where N-NO₃⁻ pool increased significantly). The introduction of NZ also affected microbial biomass N and C immobilization. In Nov-2012, lower C_mic was observed in all ZT, especially in 5NZ, suggesting that NZ introduction cause an initial disequilibrium in microbial biomass. During the sorghum-growing season (Jun-2013), microbial biomass was probably influenced by the competition with the plants, resulting in a consistent decrease in C_mic and N_mic in all the treatments. Apparently, 15NZ treatment had a stronger influence on microbial biomass N, resulting in a lower N immobilization that was already visible in May-2013 (in this case even C_mic pool was affected) and lasted until Sep-2013, suggesting more zeolitite was in the soil the less N was easily available to soil microorganisms. The C/N ratio of microbial biomass may give information regarding bacterial or fungal prevalence in the microbial community; usually bacteria have on average a lower C/N than fungi because of a lower N requirement by the latter [57,58]. Agricultural management (tillage or no-tillage), nutrient availability, soil pH, temperature and moisture may influence fungal/bacterial ratio [15,59,60]. In this respect, C_mic/N_mic ratio suggested that the introduction of natural zeolitites caused a shift in microbial population through a greater fungal prevalence [15,59,60] especially in 15NZ. A greater fungal/bacteria biomass ratio can be often interpreted as marker of a more sustainable agricultural system, where the nutrients required for the plant growth are supplied by OM decomposition and N mineralization with probably lesser N loss in the environment [15,59,60]. It is thus likely that the presence of zeolites caused a lower N accessibility into the N soil pools, making it less available to soil microorganisms and favoring the development of fungal biomass. These implications on microbial biomass affect the soil C cycle, where a lower HI and a greater FA fraction content can contribute to lower C immobilization.

4.3. Zeolitite Influence on Rhizosphere

At the harvest, rhizosphere was depleted in Exch N-NH₄⁺ with respect to bulk soils but significantly enriched in microbial biomass C-N, TOC, HA and TN. It is reasonable that the decrease in the exchangeable pool was due to an exploitation by plant roots while bulk soil results relatively enriched. Rhizosphere is known to be a zone with higher metabolic activity, thus the release of organic compounds from roots may explain the increase in microbial biomass, TOC, HA and TN [61,62] and also the slight fractionation observed in TN and Fix N-NH₄⁺ pools δ¹⁵N. It is interesting to note that at the end of the growing season, rhizosphere TOC was probably influenced by zeolitite addition (especially in 15NZ), showing an increase with respect to the CNTR. This evidence suggests a higher efficiency of microbial biomass, but this is not supported by a parallel increment in C_mic and N_mic with respect to CNTR. On the other hand, Ozbahce et al. [36] found that the application of NZ mixed in the first 30cm soil layer increased OM content, although a clear explanation of this phenomenon was not given. Italian zeolitic tuffs can adsorb a significant amount of HS from solutions, especially when divalent cations (such as Ca²⁺) act as bridge between the negatively charged mineral surface and the organic phase [63]. A plausible hypothesis is that zeolitite can adsorb HS reducing in this way organic C losses and increasing thus TOC in the soil [64]. Ca²⁺ is indeed the major exchangeable base of the studied NZ and it can act as bridge for the sorption of HS in zeolitic tuff. Further studies are required to confirm this very important aspect.

4.4. Plant-Zeolitite Interactions

N is preferentially allocated in leaves and panicles and the lack of significant differences in N content with respect to the distinct treatments suggests that, notwithstanding the fertilization reductions, plants uptake similar N amounts, reflecting a higher fertilizer recovery or N use efficiency. As reported in literature [65,66] δ¹⁵N of leaves is determined by the physiological mechanism within the plants and by the isotopic ratio of external N sources. Moreover, intra-plant N isotopic variations, between roots and leaves can be due to different patterns of N assimilation or reallocation. These intra-plant δ¹⁵N variations were recognizable in the sorghum plants, where leaves δ¹⁵N was generally greater (from 2 to 3‰) than that determined in roots, except for 7CZ where differences were even greater (about 6‰). Many authors used ¹⁵N natural abundance of leaves in order to trace the N source in the soil [65,66,67,68,69] mostly because leaves represent the primary sink of N and the largest plant N pool [41]. Average δ¹⁵N of 7CZ leaves was very close to NEZ δ¹⁵N and significantly higher than the leaves from the other parcels (Figure 6). The higher isotopic signature of 7CZ leaves indicated that during the growing season, an uptake from a significantly higher δ¹⁵N occurred. Assuming negligible fractionation effects during uptake, the main N source was probably represented by NEZ-N. Indeed, 7CZ is the only parcel where a high δ¹⁵N input (represented by pig slurry) was added during the experimentation. It has been demonstrated that crops grown under chemical fertilizers (such as urea) have lower δ¹⁵N with respect to plants grown under organic fertilizers (which have usually higher δ¹⁵N) [70,71,72]. It is important to consider that the other N inputs were urea and di-ammonium phosphate, which have a negative δ¹⁵N and were applied to all treatments, supporting the hypothesis that NEZ was an efficient slow release fertilizer and retained N even after several months and with no important N losses occurring (from leaching or denitrification). Most N in soil is bound in forms not immediately available to plants [41,71], so usually the δ¹⁵N of TN is not a good approximation of the isotopic signature of the N source preferentially used by plants. Anyhow, ¹⁵N natural abundance in leaves of CNTR, 15NZ and especially 5NZ reflect the use of N characterized by a lower δ¹⁵N and related to chemical fertilizers signature. It is important to note that notwithstanding the lower urea application (−30% with respect to the CNTR), in 5NZ and 15NZ treatments, leaves δ¹⁵N is similar or even lower with respect to CNTR. In particular, as visible from the mixing model (Figure 6), 5NZ plants showed the lowest δ¹⁵N not only in leaves but also in roots, stems and panicles, suggesting a higher efficiency in the uptake of N from a source with lower δ¹⁵N, i.e., chemical fertilizers.

The positive yield results of all ZT must be attributed also to the soil physical properties amelioration induced by zeolitite amendments. In our case, zeolitites can in fact increase soil water retention and, thanks to their coarser size with respect to the fine-grained soil (particle-size analysis performed within ZeoLIFE project are not shown in this work) they can also increase soil permeability and aeration [28,36].

5. Conclusions

In this work, a detailed investigation of the N-C pools and δ¹⁵N dynamics in the soil-sorghum system were studied in a field amended with natural and NH₄⁺-enriched zeolitites, under fertilization reductions. ZT generally had a weak influence on the total reserve of Fix N-NH₄⁺ pool. However, its reserve decreased during the growing season and a δ¹⁵N turnover occurs, confirming both its active role in plant nutrition and its dynamic equilibrium with the other soil N pools. Moreover, Fix δ¹⁵N turnover indicate that part of N from chemical fertilizers was probably stored in this pool at the end of the agronomic year. The N introduced in the soil system with NEZ increased the Exch N-NH₄⁺ reserve, suggesting that N adsorbed by zeolitites preferentially affected this pool. ZT (especially NZ) influenced the N-NO₃⁻ pool, resulting in lower nitrate contents suggesting a possible nitrification and denitrification rate reduction. Microbial activity was strongly affected by NEZ and NZ introduction which seems to cause at first a disorder, then a possible change in microbial population towards a fungal prevalence, usually indicative of a system with lower N accessible to microorganisms. ¹⁵N natural abundance in soils and plants, together with the positive yield results, suggested that plants benefited from N NEZ. However, in all likelihood, plants did not directly exploit N from NEZ, but rather the transfer involves complex interactions among soil microbial biomass. Further investigations focused on this biotic mediation are needed. Most of the N-C pools dynamics were not affected by NZ amendments but, in all likelihood, an increase in fertilizers recovery by plants occurred. The spreading of zeolitites may improve soil physical-chemical properties, enhance nutrient use efficiency as well as augment economical and an environmental saving.