Labile organic matter fractions as early-season nitrogen supply indicators in manure-amended soils

2.1 Soil sampling and amendment history

Soils were collected at the Laval University Experimental Farm near Saint-Augustin-de-Desmaures, Quebec, Canada (46°44′ N, 71°31′ W; 110 m asl). A poorly drained silty clay (432 g clay kg⁻¹, 163 g sand kg⁻¹, 35 g C kg⁻¹, pH = 6.8) classified as a mixed frigid Dystric Eutrudept and a well-drained sandy loam (170 g clay kg⁻¹, 680 g sand kg⁻¹, 19 g C kg⁻¹, pH = 7.0) classified as a mixed frigid Typic Dystrudept were selected for their contrasting properties. Experimental plots were established in May 2009 with spring wheat (Triticum aestivum L.) in a corn (Zea mays L.)–soybean (Glycine max L.)–spring wheat rotation. Above-ground crop residues were removed by hand at each harvest, and the remaining stubble was 10–15-cm above the soil surface. The contribution of crop residues to soil fertility during the crop rotation was not assessed, but should reflect the amendment history (i.e., plots with greater fertility produce higher yields and thus more crop residues).

Plots were 5 m × 7 m in size and arranged in a randomized complete block design on each soil texture with five experimental treatments [control (CTL), mineral NPK (NPK), liquid swine manure (LSM), liquid dairy cattle manure (LCM), and solid poultry litter (SPL)] replicated three times. Excluding the CTL, all plots received 90 kg available N ha⁻¹ in spring of 2009, 2010, and 2011. For the organic N sources (LSM, LCM, SPL) 90% of total N is expected to be available in the year of application for LSM, and 70% of total N is expected to be available for LCM and SPL if applied in the spring and rapidly incorporated to minimize NH₃ volatilization (CRAAQ, 2010). To simplify field work, an average coefficient of 0.8 was used for all manures, so the total N application rate for each manure was 112.5 kg total N ha⁻¹. For the NPK treatment, the mineral N fertilizer was applied as Ca-NH₄-nitrate at 90 kg N ha⁻¹. The NPK and CTL treatments received P fertilizer applied as triple superphosphate (20 kg ha⁻¹ P₂O₅ for the sandy loam; 30 kg P₂O₅ ha⁻¹ for the silty clay), and K fertilizer applied as K-chloride (20 kg K₂O ha⁻¹ for both soil types), based on local recommendations (CRAAQ, 2010). There was no amendment applied in 2012. Eight soil samples (0- to 20-cm depth) were collected with a soil auger (10 cm in diameter) from each plot in spring 2012. Samples were composited, passed through a 4.75-mm sieve while field-moist, and stored at 4°C for 14 d until the start of the experiment. Total C and N of each soil were determined by dry combustion with a Flash EA 1112 series CN analyzer (ThermoFinnigan, Italy).

2.2 Manure analysis

Composite samples from each manure source were collected on the application date in 2009, 2010, and 2011, and refrigerated (4°C) until analysis (Table 1). The LCM and LSM were homogenized using a Polytron (Model PT 3100; Kinematica AG, Littau-Lucerne, Switzerland). Distilled water was added to the SPL (10:1, water/SPL ratio) and homogenized with the Polytron to make a slurry solution. The pH of each manure was then measured by direct reading with a glass electrode. Dry matter of the manure was determined as the mass of materials remaining after drying 100 mL of LCM and LSM and 50 g of fresh SPL for 96 h at 55°C. Total C in the SPL was measured by dry combustion and total C in the LSM and LCM was measured by injecting 50 μL of homogenized into an automated combustion C analyzer (Model Formacs; Skalar Analytical, De Breda, The Netherlands). Total N and P concentrations were determined by measuring NH and PO concentrations in Kjeldahl acid digests (Chantigny et al., 2007a) with an automated continuous-flow injection colorimeter (QuickChem 8000 FIA+; Lachat Instruments, Loveland, CO, USA). The mineral N concentration of the liquid manures and SPL slurry was determined by shaking 10 mL of the sample with 50 mL of 1 M KCl for 60 min. The extract was filtered through pre-washed (1 M KCl) Whatman #42 filter paper. The NH and NO + NO concentrations in the extracts were measured with the colorimeter described above.

Table 1. Average physical and chemical characteristics (wet volume or mass basis) of the animal manures used in 2009, 2010, and 2011 for the sandy loam and silty clay soils.

Manurea	DMb	Total C	Total N		Total P	pH
	/ g L−1
LSM	71.9 ± 14.31c	30.4 ± 6.40	7.4 ± 0.77	4.7 ± 0.65	1.4 ± 0.22	7.2 ± 0.05
LCM	54.3 ± 4.72	24.0 ± 1.44	2.7 ± 0.19	1.4 ± 0.23	0.5 ± 0.04	6.6 ± 0.09
	/ g kg−1
SPL	742.2 ± 27.71	308.5 ± 12.57	27.8 ± 1.52	4.5 ± 0.28	10.1 ± 0.29	8.7 ± 0.09

• ^aLSM, liquid swine manure; LCM, liquid dairy cattle manure; SPL, solid poultry litter.
• ^bDM, dry matter.
• ^cNumbers following mean values represent SE.

2.3 Soil N supply determined from spring wheat N uptake

A short-term (42 d) plant growth study was conducted, similar to the procedure of Serna and Pomares (1991), in a controlled environment chamber (Conviron E15, Conviron Technology, Winnipeg, Canada). Day-length and temperature were set at 18 h and 18°C, and night temperature was set at 15°C. Standard 15-cm diameter pots were filled with sieved field-moist soil (850 g silty clay or 1000 g sandy loam per pot on dry soil basis) and placed on saucers. Six seeds of spring wheat (cv. A.C. Barrie) were sowed in each pot and were frequently misted with deionized water for one week until all seedlings emerged. The seedlings were then thinned to four plants per pot. Soil moisture was maintained at 55% water-filled pore space by mass measurement and watering daily. At day 14 after seedling emergence, a zero-N nutrient solution was applied based on a modified Hoagland solution (0.5 M K₂SO₄, 1 M MgSO₄, 0.05 M CaH₂PO₄) to prevent nutrient deficiency. Upon the termination of the growth trial, plant above-ground biomass was cut at the soil surface and weighed before and after oven drying at 55°C for 48 h. The soil was separated from the plant roots by gently removing the roots by hand and sieving the soil at < 4.75 mm. Plant roots were discarded as their small biomass made them difficult to salvage. Wheat root N-uptake was 16% of crop N uptake in a 45-d study with a similar cultivar and growth conditions, but was not affected by N fertilizer treatments (Thomas, 2015). Thus, discarding the wheat roots was considered to be a small source of error in this experiment. Dry above-ground plant material was ground to pass through a 1-mm screen, and the total C and N concentrations were determined by dry combustion with a Flash EA 1112 series CN analyzer (ThermoFinnigan, Italy). At the beginning (Pre-Plant) and termination (Post-Harvest) of the growth study, a 5.0 g field-moist subsample of soil was extracted with 50 mL of 2 M KCl to determine the NH and NO concentrations using the modified indophenol blue technique (Sims et al., 1995). Simultaneously, the gravimetric water content was calculated by drying 10 g of field-moist soil at 105°C for 24 h.

where Post-Harvest Mineral N is the sum of NO and NH concentrations (mg N kg⁻¹ dry soil) immediately following harvest; and Crop N uptake is the N accumulation in the above-ground biomass at harvest (mg plant N kg⁻¹ dry soil). We assumed that gaseous N losses were negligible because maintenance of 55% water-filled pore space should maximize aerobic activity and minimize denitrification (Linn and Doran, 1984). Leaching N losses were controlled because water that passed through the pot was collected in a saucer underneath the pot and poured back onto the soil surface within 24 h.

2.4 Extractable organic matter fractions

The WEOM and hot-WEOM extractions were based on the methods of Ghani et al. (2003) and Chantigny et al. (2010). Briefly, the field-moist composite soil from each experimental plot was air-dried and passed through a 4-mm sieve. A 5 g subsample was placed in a 50 mL polypropylene centrifuge tube and 30 mL of deionized water (20°C) was added. The tubes were shaken for 30 min on an oscillating shaker. The soil-water suspension was centrifuged for 20 min at 4,500 g for the sandy loam and at 10,000 g for the silty clay. The supernatant was decanted through a filter paper (Fisher Scientific Q5; pre-leached with 40 mL of deionized water) into an acid-washed Nalgene bottle. This filtered fraction was used to determine WEOC and WEON concentrations. The mass of the centrifuge tube plus the remaining wet soil was measured to calculate the entrained water volume to correct for dilution of the C and N concentrations. After filtering the supernatant, a 30 mL aliquot of deionized water (20°C) was added to each centrifuge tube and mixed thoroughly to re-suspend the soil. The tubes were tightly sealed and placed in a 50°C hot-water bath for 16 h, then tubes were centrifuged and the contents filtered again, as described above. This extraction was used to determine hot-WEOC and hot-WEON. The WEON and hot-WEON concentrations were determined after persulfate oxidation of 20°C water and 50°C hot-water extractions and were the difference between total N concentration after persulfate oxidation and the mineral N concentration of non-oxidized samples (Cabrera and Beare, 1993). The NO and NH concentrations were determined colorimetrically using the modified indophenol blue technique (Sims et al., 1995). The organic C concentration in WEOC and hot-WEOC extracts was measured with a Sievers Innovox TOC analyzer (GE Analytical Instruments, Boulder, CO, USA). The C/N ratios of WEOM and hot-WEOM were calculated by dividing the organic C concentration by the organic N concentration (mg kg⁻¹ dry soil) in each water extraction.

The POMC and POMN were measured following size fractionation (> 53 µm) based on Gregorich and Beare (2008). A 120 g subsample of field-moist soil was sieved < 2 mm and air-dried, then 25 g was dispersed in 100 mL of a weak Na-hexametaphosphate (5 g L⁻¹) solution in a 250 mL Nalgene bottle by shaking bottles sideways on a reciprocating shaker for 16 h. The POM, including the sand remaining on the 53 µm sieve was air-dried overnight and then oven dried at 50°C for 24 h and finely ground to pass through a 250-µm sieve. The concentrations of POMC and POMN were determined by dry combustion using a Flash EA 1112 series CN analyzer (ThermoFinnigan, Italy).

The MBC and MBN were measured based on Voroney et al. (2008) and Horwath and Paul (1994) using chloroform fumigation-direct extraction. A 10.0 g field-moist soil sieved < 4.75 mm was fumigated with ethanol free chloroform for 24 h, the chloroform was then repeatedly evacuated using a 80 kPa pump, and then soils were directly extracted with 0.5 M K₂SO₄. Non-fumigated soil was simultaneously extracted with 0.5 M K₂SO₄. The MBC was calculated as total organic C in fumigated samples minus the total organic C in non-fumigated samples divided by a correction factor of 0.45 (Wu et al., 1990; Joergensen, 1996). The organic C concentrations were determined by a Sievers Innovox TOC analyzer (GE Analytical Instruments, Boulder, CO, USA). The MBN was calculated as total N in fumigated samples minus the total N in non-fumigated samples. For fumigated and non-fumigated soils, the MBN was determined by subtracting the NO concentration of non-oxidized samples from the NO concentration of persulfate oxidized samples (Cabrera and Beare, 1993). The NO concentrations were determined colorimetrically using the modified indophenol blue technique (Sims et al., 1995). The MBN was then divided by a correction factor of 0.54 (Joergensen and Mueller, 1996). The concentration of organic C in the non-fumigated 0.5 M K₂SO₄ extract was considered to represent the SEOC. The SEON concentration was calculated as the difference in total NO concentration of non-fumigated 0.5 M K₂SO₄ persulfate oxidized and non-oxidized samples (Chantigny et al., 2007b).

2.5 Statistical analysis

All data were first tested for normality using the Shapiro–Wilks W test, and log-transformed as required. The effect of fertilizer treatments on the WEOC, WEON, hot-WEOC, hot-WEON, POMC, POMN, MBC, MBN, SEOC, SEON concentrations, on the C/N ratio of each fraction, and SNS was analyzed by one-way ANOVA for each soil texture using the PROC GLM procedure of SAS 9.3 software (SAS Institute, 2011; Cary, NC, USA). When there was a significant fertilizer treatment effect (P < 0.05), differences among treatment means were evaluated using the least significant difference (LSD) test with TUKEYs adjustment at a P < 0.05 confidence level. Correlations among these parameters were determined with Pearson correlation coefficients (r) using the PROC CORR statement in SAS 9.3 software using all (n = 30) treatment replicates. In addition, simple linear regression analyses were performed for the POMN and POMC, and SNS for each treatment replicates (n = 30) using the PROC REG statement in SAS 9.3 software. Significant effects were accepted at a P < 0.05 confidence level.

3.1 Soil N supply

Since the SNS was not affected by fertilizer treatments in soils with contrasting texture (Table 2), data were pooled among soil types to evaluate the relationships between labile OM fractions and the SNS. The strongest correlations were between the SNS and the POMN concentration (r = 0.87; P < 0.001; Table 4, Fig. 1), the POMC concentration (r = 0.83; P < 0.001), the SOC concentration (r = 0.82, P < 0.001), and the hot-WEON concentration (r = 0.81, P < 0.001). When the C/N ratios of labile OM fractions were correlated with SNS, the correlation coefficients were lower than those of the individual components (i.e., POMN > POMC:POMN), suggesting a stronger association between POMN and other measured parameters to the SNS than the C/N ratio of any labile OM fraction (Table 4).

4.1 Labile OM concentrations

Fertilized soils in this study exhibited a range of C and N concentrations in labile OM fractions. The WEON was the only fraction that supported the hypothesis of higher labile organic N concentration in SPL-amended soils than other fertilizer treatments, implying that SPL increased the WEON concentration. However, there was more hot-WEON and SEON in the LCM-amended silty clay soil than the SPL treatment, which suggests that 50°C water extraction and salt extraction procedures retrieve potentially soluble substances such as microbial by-products (Courtier-Murias et al., 2013) that cannot be extracted with 20°C water. Lack of difference among fertilizer treatments was not unexpected because soil samples were collected in the spring after the field plots—which were not planted with a cover crop after wheat harvest—were subjected to 7 months of high precipitation (rainfall and snow), freeze–thaw cycles during winter and spring snowmelt. In this humid temperate climate, such conditions are expected to result in N losses through runoff, leaching, and in gaseous emissions (Rasouli et al., 2014), which appears to deplete the C and N concentrations in labile OM fractions towards baseline levels, regardless of the fertilizer history.

4.2 Labile OM as an indicator of SNS

In a humid temperate climate, it is considered practical to perform soil test analyses for N at pre-planting (e.g., in the week before the field is seeded) to predict the SNS to field crops during vegetative growth stages when crop N demands are high. This was the rationale for measuring labile OM fractions prior to beginning the controlled environment study with wheat. Although the SNS after 42 d was not affected by fertilizer treatments, the variation in labile OM fractions across 5 fertilizer treatments in two soil textures permitted correlation analysis. The best indicators of the SNS were the POMN and POMC concentrations, which is consistent with our hypothesis, but the SOC content in the whole soil was also strongly correlated with the soil N supply and slightly better than the hot-WEON concentration. It seems logical that soils with higher SOC content also possess greater POMN, POMC, and other labile OM fractions. The question is whether we should be measuring the labile OM fractions to predict SNS or whether whole soil measurements provide insight into the SNS.

The fact that POMN was the strongest indicator of the SNS in this study is consistent with findings in soils amended with crop residues from W Canada (St. Luce et al., 2011), E Canada (St. Luce et al., 2014) and a loam soil with a 13-y history of solid beef cattle manure in E Canada (Sharifi et al., 2008). Since the top two indicators of SNS were based on the POM fraction, it suggests that the biological processes that are responsible for producing mineral N are controlled by POM. Two distinct possibilities emerge: (1) the POM fraction is a source of organic N that is transformed into mineral N; and (2) the POM fraction is an accessible organic C substrate for microbial biomass that controls N mineralization from other sources (e.g., WEON). The first possibility may be supported in the silty clay soil, which was a N-rich substrate with a low POMC:POMN ratio of 11 to 12, larger microbial biomass C and N and a greater SNS than the sandy loam soil. On the other hand, the POMC:POMN ratio ≥ 25 in the sandy loam soil implies that POM is primarily a C substrate and so N mineralization is likely occurring from other labile OM fractions. Given the rapid turnover of soluble OM fractions, the hot-WEON fraction warrants further consideration as an indicator of SNS.