Fate of nitrogen derived from ¹⁵N-labeled cattle manure compost applied to a paddy field in the cool climate region of Japan

Abstract

To estimate the fate of nitrogen (N) derived from cattle manure compost with sawdust (CMC) in a paddy field in the cool climate region of Japan, well-composted ¹⁵N-labeled CMC was applied to a microplot field experiment. Throughout the experimental period of three crop seasons, N from CMC was taken up by rice plants without a marked decline. The percentages of N taken up derived from CMC to applied N as CMC (%CNRp) were 2–3% for each year. The N from CMC was taken up by rice plants over the entire growth period by 1–2, 2 and 2–3% as %CNRp at the panicle initiation, heading and maturity stages, respectively. A significant positive linear correlation was found between the cumulative compost N uptake and the number of days transformed to standard temperature (25°C) over the entire experimental period, including the fallow season. The %CNRp was identical at CMC application rates ranging from 1 to 4 kg m⁻². Using ¹⁵N-labeled CMC, the results showed that well-composted CMC was a stable N source for rice plants for at least 3 years, regardless of the CMC application rate (ranging from 1 to 4 kg m⁻²) in the cool climate region of Japan. The distribution of CMC N was 7% in the rice plants accumulated over 3 years, 66–69% in the soil and 24–27% was un-recovered at the end of the third crop season.

Key words:

• cattle manure
• compost
• ¹⁵N
• paddy field
• rice

INTRODUCTION

The demand for agricultural produce, including rice, grown using organic materials as a nutrient source is increasing with recognition of the importance of resource recycling farming systems (The Ministry of Agriculture, Forestry and Fisheries of Japan 2006). At the same time, the cultivation area of forage rice is expanding, where all the aerial parts of the rice plant are removed from the paddy fields (Agriculture, Forestry and Fisheries Research Council 2006). In paddy fields where forage rice is cultivated, the application of organic materials is essential to avoid deterioration of soil fertility. In contrast, in intensive livestock-farming areas, a great amount of nitrogen (N) is excreted through livestock waste (Ikumo 2003; Tsuiki and Harada 1997). Estimation of the N balance indicates that surplus excreta N is loaded in some of these areas (Kohyama et al. 2003), which necessitates the effective use of livestock waste in areas other than the intensive livestock-farming areas. Thus, the effective use of organic materials, livestock manure compost in particular, in rice farming is anticipated. For the effective use of livestock manure compost in a sustainable recycling system, information about the fate of livestock manure compost N in the paddy field is indispensable. To accurately evaluate the fate of N in organic materials, the use of ¹⁵N-labeled organic materials is a relevant approach (Hood et al. 1999; Muñoz et al. 2003; Takahashi et al. 2000). The fate of N from ¹⁵N-labeled livestock manure in paddy fields has mostly been reported from the warm climate regions of Japan (Matsushita et al. 2000a, b; Matsuyama et al. 2003; Nishida and Tsuchiya 2001; Nishida et al. 2004, 2005; Ueno and Yamamuro 2001; Uenosono et al. 2004; Yamamuro et al. 2002). However, this methodology has not been applied in cool climate regions. Hence, the fate of livestock manure N in paddy fields remains vague under cool climate conditions in Japan. Among previous reports, only two studies have covered multiple-year monitoring of applied ¹⁵N (Matsuyama et al. 2003; Ueno and Yamamuro 2001). Rice plants took up livestock manure ¹⁵N over multiple years in these studies, which shows the significance of multiple-year observations of organic material N. The objective of this study was to directly evaluate the fate of N originated from cattle manure compost over 3 years by applying ¹⁵N-labeled compost to a paddy field in the cool climate region of Japan.

MATERIALS AND METHODS

Field experiment

The microplot experiment inside a paddy field was conducted from 2000 to 2002 at the National Agricultural Research Center for Tohoku Region (NARCT), Daisen, Akita, Japan (N39°29′, E140°30′, altitude 30 m). The soil is a fine-textured gray lowland soil (Typic Fluvaquents; Soil Survey Staff 1998) and soil in the plow layer contained 2.35 g kg⁻¹ total N on a dry-weight basis. A polyvinyl chloride frame (17 cm × 30 cm, height 20 cm) wrapped at the bottom with unwoven cloth (Unisel BT-1808W; Unisel, Iwakuni, Japan) was installed into the plow layer (approximately 15 cm depth) inside the paddy field on 22 May 2000. Fresh soil previously collected from the plow layer (6.27 kg dry weight) was passed through a 1-cm sieve and was well mixed with ¹⁵N-labeled cattle manure compost that contained sawdust (CMC). Compound fertilizer (Takihosuka 3 gou; Taki Chemical Company, Kakogawa, Japan) was also mixed with the soil. The mixture of soil, CMC and fertilizer was put into the flame on 22 May 2000. The area inside the flame was referred to as a microplot in this study.

The application rates of CMC were 1, 2 and 4 kg m⁻² on a fresh-weight basis, and the rate of compound fertilizer was 8 g m⁻² (as N, P₂O₅ and K₂O). The ¹⁵N-labeled CMC was made by Yamamuro (Osaka, Japan) using feces collected from a cow fed with ¹⁵N-labeled corn (Zea mays L.). Before composting, sawdust was added to the cattle feces at a rate of approximately 30% to make suitable moisture conditions (Yamamuro 2000).

Puddling was carried out on 23 May. On 25 May, 35-day-old rice plant seedlings (Oryza sativa L., cv. Akitakomachi), three seedlings per microplot, were transplanted at the center of each microplot. A hole (2-cm diameter) in each side of the frame, slightly higher than ground level, was sealed with a rubber plug within a few days after puddling and was opened thereafter to allow ponding water to pass through. To the field outside the microplots, only compound fertilizer was applied at a rate of 8 g m⁻² (as N, P₂O₅, and K₂O), and rice plants were planted at a spacing of 17 cm × 30 cm, the same planting density as the microplot.

Three replicates were set up for the 1 and 4 kg m⁻² CMC treatments. For the 2 kg m⁻² treatment, 21 replicates were initially set up because 18 of 21 replicates were to be eliminated at on-season sampling times (panicle initiation and heading stages) for 3 years as described later. The arrangement of microplots followed a completely randomized design in three rows. The interval of each row, in which microplots were placed, was 90 cm and two rows of rice plants were placed there. The interval of each microplot within a row (center to center) was 51 cm and two hills were placed there.

Three replicate plant top samples in the 2 kg m⁻² CMC treatment were collected at the panicle initiation stage (10 July) and at the heading stage (3 August). Microplots where plant samples were collected at the panicle initiation or heading stages were eliminated from the experiment after that time. Plant top and soil samples of all other microplots, three replicates for the 1 and 4 kg m⁻² CMC treatments, and 15 replicates for the 2 kg m⁻² treatment, were collected at maturity (13 September). Plant roots were left in the soil. After sample collection at the maturity stage, the microplots (soil and root in the frame) remained in the ground, up to approximately 15 cm depth, until the next crop season.

In 2001, only compound fertilizer was applied to the soil of the microplots on 21 May following the same rate as in the previous year. Puddling and transplanting were carried out on 22 May and 27 May, respectively. The method of arrangement of microplots was the same as that in 2000 (randomized in three rows). Three replicate plant top samples in the 2 kg m⁻² CMC treatment were collected at the panicle initiation stage (10 July) and at the heading stage (7 August). The microplots where plant samples were collected at the panicle initiation or heading stages were eliminated from the experiment after that time. Plant top and soil samples from all other microplots, three replicates for the 1 and 4 kg m⁻² CMC treatments, and nine replicates for the 2 kg m⁻² treatment, were collected at the maturity stage (18 September). Plant roots were left in the soil. After sample collection at the maturity stage, the microplots (soil and root in the frame) remained in the ground, up to approximately 15 cm depth, until the next crop season.

Similar procedures were carried out in 2002. Only compound fertilizer was applied at the same rate used in 2000 and 2001. Puddling and transplanting were carried out on 24 May and 29 May, respectively. Three replicate plant top samples in the 2 kg m⁻² CMC treatment were collected at the panicle initiation stage (15 July) and at the heading stage (12 August). The microplots where plant samples were collected at the panicle initiation or heading stages were eliminated from the experiment after that time. Three replicate plant top and soil samples from all other microplots were collected at the maturity stage (25 September).

Table 1 Chemical properties and biological oxygen consumption of 15N-labeled cattle manure compost that contained sawdust

N (g kg^−1)	^15N (atom%)	C (g kg^−1)	C/N ratio	NH4-N (g kg^−1)	NO3-N (g kg^−1)	Moisture (g g^−1)	Biological oxygen consumption (µg g^−1 min^−1)
3.95	2.54	93.0	24	0.15	0.02	0.73	1
Fresh-weight basis.

Sample preparation and analysis

Plant samples were well washed and oven-dried at 70°C. Plant samples were then weighed and ground using a Willey-type cutting mill (Fujiwara Scientific Company, Tokyo, Japan). Soil samples were air-dried and passed through a 2-mm sieve. The soil and plant samples were digested by sulfuric acid with the occasional addition of hydrogen peroxide, and then the digest was diluted to 100 mL in a measuring flask. An aliquot of this solution was subjected to the Kjeldahl distillation method for total N analysis. From another aliquot, a solution containing 100 µg N was put into a 50-mL Erlenmeyer flask to condense its N to a form of NH₄Cl by HCl following the diffusion method (Yoneyama et al. 1975). This NH₄Cl solution was transferred to a tin capsule and subjected to ¹⁵N analysis using Automated Nitrogen and Carbon Analyzer-Solid and Liquid (ANCA-SL; PDZ Europe, Cheshire, UK). Total N and ¹⁵N abundance of the compost were measured following the same procedure. Water extractable NH₄-N and NO₃-N of the compost were measured with an Aquatec 5400 Analyzer (Foss Tecator, Hilleroed, Denmark) after extraction by distilled water in a ratio of compost (fresh weight [FW]) : distilled water of 1:10 for 1 h. Total carbon (C) of the compost was measured with a Vario MAX CN (Elementar Analysensysteme, Hanau, Germany). The biological oxygen consumption of the compost was measured for 50 g of the compost (FW) for 30 min at 35°C using a Compotester (Fujihira Industry, Tokyo, Japan).

Calculation of the number of days transformed to standard temperature (25°C)

Using daily mean air temperatures during the experimental period, including the fallow season (National Agricultural Research Center for Tohoku Region 2001, 2002, 2003), the number of days transformed to a standard temperature of 25°C (DTS) was calculated using the following equation, which was obtained from Arrhenius’ law (Sugihara et al. 1986),

where T_a was the daily mean air temperature (K), T_s was the standard temperature (298 K), R was a gas constant (1.987 cal deg⁻¹ mol⁻¹) and Ea was the apparent activation energy (cal mol⁻¹). As an Ea value, 13,800 kcal mol⁻¹ was used, which was obtained by an incubation experiment of CMC under submerged conditions (Sakai and Yamamoto 1999). The relationship between DTS and N uptake derived from CMC (Ndfc) in the 2 kg m⁻² CMC treatment was examined using a single regression analysis.

Calculation of the relative efficiency of compost N to chemical fertilizer N

The relative efficiency of compost N to chemical fertilizer N, which is generally used as an index of compost N efficiency, was calculated using the following equation:

where %CNRp was the percentage of compost N recovery by the rice plants and %FNRp was the percentage of chemical fertilizer N recovery by the rice plants (Nishio 2007). Forty percent was used as the %FNRp value, which was observed by applying (¹⁵NH₄)₂SO₄ as a basal dressing in the same field for the same variety of rice (M. Nishida, unpubl. data, 2004–2006).

Statistical analysis

A Tukey–Kramer multiple comparison was carried out on the data to clarify the effect of CMC application rate. A single regression analysis was conducted on cumulative Ndfc in the 2 kg m⁻² CMC treatment and the DTS over the entire experimental period, including the fallow season. Statistical procedures were carried out using JMP (SAS Institute 2002).

RESULTS AND DISCUSSION

Maturity of the CMC

The chemical properties and biological oxygen consumption of the CMC are listed in Table 1. The compost contained a small amount of inorganic N and was almost odorless. Biological oxygen consumption was as low as 1 µg g⁻¹ min⁻¹. In cases where the biological oxygen consumption of compost is less than 3 µg g⁻¹ min⁻¹, the compost can be considered to be in a stable phase with high maturity (Furuya et al. 2003). Therefore, the CMC used in this study can be regarded as highly matured CMC.

Table 2 Total N uptake of rice plants, N derived from 15N-labeled cattle manure compost with sawdust (CMC) in rice plants (Ndfc), and the percentage of Ndfc to applied N as CMC (%CNRp)

	Year	CMC application rate (kg m^−2)	Growth stage
			Panicle initiation	Heading	Maturity
Total N uptake (g m^−2)	2000	1	−	−	9.51 ± 0.26 b
		2	6.63 ± 0.34	8.19 ± 0.34	9.80 ± 0.42 b
		4	−	−	10.74 ± 1.01 a
	2001	1	−	−	8.52 ± 0.34 a
		2	6.16 ± 0.20	7.90 ± 0.38	8.67 ± 0.20 a
		4	−	−	8.85 ± 0.45 a
	2002	1	−	−	9.00 ± 0.52 a
		2	5.98 ± 0.16	8.15 ± 0.05	9.02 ± 0.34 a
		4	−	−	9.24 ± 0.38 a
Ndfc (g m^−2)	2000	1	−	−	0.10 ± 0.02 c
		2	0.12 ± 0.03	0.17 ± 0.03	0.21 ± 0.02 b
		4	−	−	0.45 ± 0.00 a
	2001	1	−	−	0.08 ± 0.00 c
		2	0.08 ± 0.01	0.12 ± 0.01	0.16 ± 0.01 b
		4	−	−	0.31 ± 0.02 a
	2002	1	−	−	0.09 ± 0.01 c
		2	0.08 ± 0.01	0.16 ± 0.01	0.19 ± 0.02 b
		4	−	−	0.39 ± 0.02 a
%CNRp (%)	2000	1	−	−	2.6 ± 0.6 a
		2	1.5 ± 0.3	2.1 ± 0.4	2.6 ± 0.3 a
		4	−	−	2.8 ± 0.0 a
	2001	1	−	−	1.9 ± 0.1 a
		2	1.0 ± 0.1	1.6 ± 0.2	2.0 ± 0.2 a
		4	−	−	2.0 ± 0.2 a
	2002	1	−	−	2.4 ± 0.2 a
		2	1.1 ± 0.1	2.1 ± 0.1	2.4 ± 0.2 a
		4	−	−	2.5 ± 0.1 a
Mean ± standard deviation (n = 15 for the 2 kg m^−2 CMC treatment at the maturity stage in 2000, n = 9 for the 2 kg m^−2 CMC treatment at the maturity stage in 2001, and n = 3 for the remaining treatments). Means followed by the same letters within a year are not significantly different using a Tukey–Kramer test (P < 0.05).

Nitrogen efficiency of CMC for rice plants

Total N uptake, Ndfc (N derived from CMC in rice plants) and the percentage of Ndfc to applied N as CMC (%CNRp) are listed in Table 2. Total N uptake was highest in 2000 followed by 2002, and was lowest in 2001. The N originated from CMC was taken up by the rice plants in the three crop seasons. The Ndfc and %CNRp showed similar trends to total N uptake with respect to yearly changes. The %CNRp values at the maturity stage were 2–3% for each year, and no marked decline in %CNRp was observed. Hence, N efficiency of the CMC was stable up to the third crop season from CMC application. In the 2 kg m⁻² CMC treatment, N uptake from CMC continued throughout the growth period for the three crop seasons, and was 1–2, 2 and 2–3% as %CNRp at the panicle initiation, heading and maturity stages, respectively. Thus, using ¹⁵N-labeled CMC, the results showed that CMC with high maturity was a stable N source for rice plants throughout the entire 3-year growth period.

The relationship between DTS (number of days transformed to standard temperature at 25°C) and cumulative Ndfc over the 3 years in the 2 kg m⁻² CMC treatment is shown in Fig. 1. A significant positive linear correlation was found between the DTS and the cumulative Ndfc, suggesting that mineralization of CMC N and its uptake by rice plants depended on the temperature. Increases in Ndfc between the panicle initiation and heading stages were greater than those indicated by the regression line. Inorganic N released from the CMC during the early growth stages of the rice plants would not be immediately taken up, and part of the inorganic N could remain in the soil. After the rice plants had grown to some level, the inorganic N, including this remaining N, would be rapidly taken up. This would result in a greater increase in the Ndfc between the panicle initiation and heading stages than that observed in the regression line.

Table 3 Nitrogen distribution of 15N-labeled cattle manure compost that contained sawdust

	Distribution (%)^†			C/N
	Rice plants	Soil	Unrecovered
Ueno and Yamamuro (2001)	12	33	55	No data
Nishida et al. (2004)	6	62	32	31
Uenosono et al. (2004)	4	91	5	37
This study	3	73–78	20–25	24
^†Distribution after one crop season after application.

Figure 1  Relationship between the number of days transformed to a standard temperature (25°C) (DTS) over the entire experimental period, including the fallow season, and the cumulative N uptake originated from 15N-labeled cattle manure compost with sawdust (CMC) in the 2 kg m−2 CMC treatment. Ndfc, N uptake originated from CMC. ***P < 0.001.

Total N uptake increased with CMC application rate, although no significant difference was observed in 2001 and 2002 (Table 2). The Ndfc at the maturity stage significantly increased with CMC application rate, and was approximately 0.1, 0.2 and 0.3–0.5 g m⁻² in the 1, 2 and 4 kg m⁻² CMC treatments, respectively. The %CNRp at the maturity stage was almost constant regardless of the application rate of CMC, indicating N efficiency per unit weight of the CMC was similar at application rates ranging from 1 to 4 kg m⁻².

Comparison of yearly changes in N efficiency of CMC with reported data

The %CNRp values in a previous study conducted at Fukuyama, Hiroshima, were 12.3, 2.9 and 0.7% in the first, second and third years, respectively (Ueno and Yamamuro 2001; values in the second and third years were calculated using a first-order kinetics regression). Similarly, the %CNRp values in Kasai, Hyogo, were 11.0, 3.0, 4.3, 3.7 and 0.4% in the first to fifth years, respectively (Matsuyama et al. 2003). In both reports, the uptake of N from ¹⁵N-labeled CMC with sawdust was observed in fine-textured gray lowland soil over multiple years. In Ueno and Yamamuro (2001), the %CNRp drastically declined and became very low in the third year. In Matsuyama et al. (2003), the %CNRp in the first year appeared to be the highest and decreased in the second year. For 3 years from the second to fourth years, the %CNRp kept a similar level, and dropped down to a very low value in the fifth year. The result of the present study was distinctive, in that the %CNRp remained at a similar level for the 3 years.

One of the possible reasons for this discrepancy is a difference in the properties of the CMC. Higher %CNRp in the first year, observed in the previous studies, might be attributed to higher contents of inorganic N and/or N in easily decomposable organic matter in the CMC, although these contents are unknown. In Matsuyama et al. (2003), however, similarity can be found with the present study in that stable %CNRp values were observed for multiple years. In the case of the CMC used in Ueno and Yamamuro (2001), the portion of the compost N that remained in the soil was much lower than that of the present study after the first crop season (Table 3). It can be speculated that a high portion of the compost N used in Ueno and Yamamuro (2001) was very vulnerable.

In addition, a higher temperature might promote the decomposition of organic matter in warmer climatic conditions. As seen in the present study, Ndfc is related to temperature. The average values of the annual mean air temperature, mean values from 1971 to 2000, in Hiroshima and Hyogo (Japan Meteorological Agency 2007) are approximately 5°C higher than that in Daisen (National Agricultural Research Center for Tohoku Region 2006). In the other reports, which showed single-year N recovery from CMC by rice plants in warmer climatic conditions, the %CNRp values were 4–9% and higher than those found in the present study (Matsushita et al. 2000a,b; Nishida et al. 2004; Uenosono et al. 2004).

Relative efficiency of CMC to chemical fertilizer

The relative N efficiency of CMC to chemical fertilizer is generally recognized to be approximately 30% (Harada 1998). However, the relative efficiency of the CMC used in the present study was estimated to be 7% for the applied year, which was lower than that generally expected. In previous studies conducted at Fukuoka (Nishida et al. 2004) and Kagoshima (Uenosono et al. 2004), the relative efficiencies of ¹⁵N-labeled CMC were 16–19% and 10%, respectively, which are also lower than 30%. These results indicate that the relative efficiency of CMC can be lower than that generally expected; implying that the relative efficiency of CMC should be scrutinized using a ¹⁵N tracer technique for various CMC, for example, different additives and degrees of maturity under diverse conditions.

Distribution of CMC nitrogen

The distribution of CMC N was 6.9–7.3% in the rice plants, 66–69% in the soil and 24–27% was unrecovered at the end of the third crop season (Fig. 2). The percentage of compost N distributed to the soil did not significantly differ with CMC application rate throughout the experimental period (the results of the multiple comparison analysis are not shown in Fig. 2). The distributions of compost N were similar in all application rates of CMC. The greatest portion of compost N, approximately 70%, remained in the soil even after the third crop season. This remaining N would continue to be a N source for rice plants.

The unrecovered part could represent the loss of the compost N and the N derived from CMC in a substantial part of the rice roots that did not pass through a 2-mm sieve. Nitrogen in the roots was not measured in the present study. However, taking into account the ratios of top to root weight of rice plants, which were 5–20 (Nagoshi et al. 2001; Terashima et al. 1994), most of the unrecovered part could be ascribed to the loss. The greatest loss was found in the first crop season. The reason for this is unknown. Although the tested CMC in our study was highly matured, inherent inorganic N and N in easily decomposable organic matter in the CMC might be lost in the early growth period, before the root system is well developed, in the first crop season. A limited number of studies have demonstrated the N distribution of CMC in paddy fields (Table 3). Among these studies, including the present study, the balances of compost N distribution are different, and a general explanation about N distribution of CMC cannot be made at this time. The amount of water percolation in a paddy field in the NARCT was 10 mm day⁻¹ (Sumida 1993). Percolation in the field could affect the loss of compost N by leaching. In this study, however, loss through leaching, denitrification and volatilization were not separately evaluated. Hence, the contribution of the loss through leaching to all loss is unclear.

Figure 2  Distribution of N originated from 15N-labeled cattle manure compost with sawdust (CMC) at the maturity stage. Error bars indicate standard deviation (n = 15 for the 2 kg m−2 CMC treatment in 2000, n = 9 for the 2 kg m−2 CMC treatment in 2001 and n = 3 for the remaining treatments). %CNRp, percentage of N taken up derived from CMC to applied N as CMC.

Conclusion

Using ¹⁵N-labeled CMC, characteristics regarding the fate of N derived from well-composted CMC, which was applied to a paddy field in a cool climate region, appear to be: (1) N originated from CMC was taken up by rice plants for at least 3 years without marked decline, (2) within a crop season, the N from CMC was taken up over the entire growth period, (3) cumulative Ndfc linearly increased with DTS for 3 years including the fallow season, (4) at application rates ranging from 1 to 4 kg m⁻², the N uptake efficiency per unit weight of CMC was identical, (5) the relative N efficiency of the CMC to chemical fertilizer was lower than that recognized generally, (6) approximately 70% of compost N remained in the soil even after the third crop season.

As a greater portion of compost N remained in the soil even after 3 years, longer observations need to be carried out. In addition, compost is often applied successively in practical farming. It is, therefore, necessary to quantify the effect of successively applied compost N on rice plants, paddy soil and the environment using ¹⁵N-labeled compost.

ACKNOWLEDGMENTS

We thank the Akita Agricultural Experiment Station for cooperation with the ¹⁵N analysis, and Dr M. C. Casimero, Dr H. Sekiya and Dr K. Yoshida for helpful suggestions.

Notes

Present addresses: National Agriculture and Food Research Organization, Tsukuba, Ibaraki 305-8517, Japan.

National Agricultural Research Center, Tsukuba, Ibaraki 305-8666, Japan.