T. TAKAHASHI, Headquarters of the National Agriculture and Food Research Organization, 3-1-1, Kannondai, Tsukuba City, Ibaraki 305-8517, Japan. Email: firstname.lastname@example.org
Denitrification of paddy fields is a key process for improving water quality in fields where nitrate concentrations are high. The objective of the present study was to understand the effects of incorporating organic carbon (C) into soil on the denitrification rate of paddy fields in winter. On 11 December 2007, separate paddy field plots were prepared by incorporating 5 Mg ha−1 of rice straw (RS), 11 Mg ha−1 of rice straw compost (RSC) or a control. A field with a high concentration of nitrate in the water (averaging 18 mg N L−1) was irrigated until 29 March. During the experiment, the daily average soil temperature at a depth of 0.05 m ranged between 3 and 15°C. The nitrate concentration in the surface water in the RS plot, where the residence time was 2 days, decreased more than the concentration in the control or RSC plots. The total estimated nitrate removal from each plot in relation to the other plots was RS > RSC = control. Measurements of the soil from each plot on 29 February 2008 showed that incorporation of RS significantly increased the denitrification potential, even at low temperatures (5–10°C). Furthermore, the RS plot contained more dissolved organic C than the control or RSC plots. This result indicates that supplying RS effectively increases denitrification under low-temperature conditions.
The potentially adverse effects of fertilizer-derived NO3− on environmental quality and public health are well known. Nitrate removal by denitrification is one of the most effective techniques for removing the nitrate load from areas used for intensive agriculture (Tabuchi et al. 2001). A typical case is provided by efforts to remove nitrate discharge from tea fields, which receive high amounts of nitrogen fertilizer (Nira and Atsumi 2007). Because tea fields received more than 1,000 kg N ha−1 year−1 of nitrogen fertilizer, high nitrate concentrations are sometimes observed in streams near tea fields and in the watersheds containing the streams. Recently, Nira and Atsumi (2007) studied nitrate removal from paddy fields located in the watersheds of such streams. They found that annual irrigation of paddy fields with stream water removed nitrate from the streams, even if the nitrate concentration of the stream was more than 15 mg N L−1. The results imply that extending the period of paddy field irrigation from the conventional mid-spring through summer period to a winter through summer period would be useful to reduce the nitrate load on inland intensive agriculture areas. The study pointed out that the nitrate concentration in the surface water does not effectively decrease in the winter; however, the concentration in the percolated water does decrease. This is caused by low temperatures, which decrease the microbial activity related to denitrification. Greater nitrate removal from surface water under low-temperature conditions is preferable for achieving stable nitrate removal in fields where permeability is relatively low.
Organic carbon (C), which is utilized by denitrifying heterotrophic microorganisms that reduce nitrate, is a key factor in the control of the denitrification rate. Starr and Gillham (1993) suggested that the addition of organic C to an environment that lacks C sources could increase the denitrification rate. If the enrichment of C sources increases the denitrification rate at low temperatures, a combination of an extended irrigation period and incorporation of organic matter could increase nitrate removal. However, to date there have been no studies to suggest that the incorporation of C sources increases nitrate removal under low-temperature conditions. The objective of the present study was to examine the effect of the incorporation of organic matter on nitrogen removal under such conditions.
Materials and methods
The field experiment was conducted at a paddy field in Kami-Kurasawa, Kikugawa city, Shizuoka prefecture, Japan. The soil was classified as Typic Endaquepts. There was one control plot and two plots into which one type of organic matter was incorporated as a C source, either rice straw (RS) or rice straw compost (RSC). The amount of RS incorporated was based on the amount of rice straw obtained per unit area. Based on fresh weight, 5 Mg ha−1 of RS can be achieved and 11 Mg of RSC can be made from 5 Mg of RS; therefore, the amounts incorporated into the plots were 5 Mg ha−1 RS and 11 Mg ha−1 RSC. The RS and RSC were imported from the Shizuoka Research Institute of Agriculture and Forestry. The RSC was prepared by composting RS for 120 days with 5% slaked lime and 0.4% nitrogen from urea. The amounts of incorporated organic matter, C and nitrogen are shown in Table 1. Each plot measured 15 m × 2 m and was separated from the other plots with plastic walls inserted around the perimeter to prevent cross-contamination of the surface water. The RS and RSC were incorporated on 11 December 2007. The plots were tilled to a depth of 0.13 m on the same day using a rotary tiller. Eight days after the incorporation, water containing a high nitrate concentration, ranging from 13 to 23 mg N L−1, was used to irrigate each plot at a rate of 1 L min−1. The irrigation water was taken from a stream near the plots. The high nitrate concentration of the water was caused by fertilizer applied in the area upland of the paddy fields (Matsuo and Nonaka 2002). The soil temperature of each plot was measured at a depth of 0.05 m. In each plot, porous cups were inserted at both 0.1 m and 0.2 m to measure the nitrate concentration in the water that percolated from the soil. The plow pan of the field was 0.15 m below ground level.
Table 1. Amounts of incorporated organic matter, carbon and nitrogen
Rice straw compost
DW, dry weight.
kg DW ha−1
The nitrate concentration of both the surface and percolated soil water was measured every 10 days. Because the residence time of the surface water in the plots was too short to estimate the nitrate removal of the entire water system of the paddy fields, the dead surface water was measured on each measuring day, to limit the residence time to 2 days. A 2-day residence time is the annual average for paddy fields where the experiment was conducted; the areas of the paddy fields, water storage and inflow were 8,200 m2, 410 m3 and 221 m3day−1, respectively. Two polyethylene tubes with a diameter of 0.1 m were inserted down into the plow pan and approximately 5 mL of surface water was sampled to obtain the initial nitrate concentration of the surface water. Two days later, the surface water in the tube was sampled again. The nitrate removal rate was calculated as the decrease in the nitrate concentration calculated from the two samplings. Nitrate concentrations were determined by ion chromatography (DX-100; Dionex, Sunnyvale, CA, USA). The nitrate concentrations of the percolated water were obtained from soil water collected in the porous cups. Measurements were continued until 29 March 2008. All measurements, except for the nitrate concentrations in the percolated water, were conducted in duplicate. Significant differences in the nitrate concentrations between the treatments were statistically analyzed using a paired t-test. That is, the differences in each treatment were evaluated throughout all measuring periods, with the nitrate concentration on each measuring day paired. The depth of floodwater was also measured at each measuring time with a gauge. The averages before and after inserting the tubes were defined as the depth of floodwater at the measuring time.
Estimation of soil denitrification potential and dissolved organic carbon
Surface soils of <0.13 m depth were sampled at each plot on 29 February 2008 and the denitrification potential and dissolved organic carbon (DOC) were measured. The denitrification potential was determined using the acetylene block method of Tsunekawa et al. (2006). Two milliliters of soil paste was placed in a 30-mL vial and sealed with a butyl rubber cap. The headspace air was replaced with N2 gas (purity >0.99999), and 5 mL of 5 mmol L−1 KNO3 and 3 mL of acetylene gas were added. The samples were incubated at 5, 10 or 15°C for 4 h. Next, 1 mL of chloroform was added to stop further denitrification. The N2O concentrations in the headspaces of the vials were measured by gas chromatography with an electron capture dissociation (ECD) detector (AC-2010; Shimadzu, Kyoto, Japan). The denitrification potentials were calculated by subtracting the amount of N2O with acetylene addition from that without acetylene addition. For DOC measurements, 10 mL of soil paste and 50 mL of distilled water were placed in a 100-mL tube, the tube was shaken for 1 h, and the supernatant was filtered through a 47-μm membrane filter. The DOC of these solutions was analyzed using a total organic carbon analyzer (TOC-5000A; Shimadzu). All measurements were conducted in triplicate. The differences between the denitrification potential and the DOC of the control and each treatment data point (RS and RSC incorporation) were statistically analyzed using a t-test.
Results and Discussion
The nitrate concentrations of the irrigated water and the soil temperatures are shown in Figures 1 and 2, respectively. Almost every day during January and February, the soil temperature was below 5°C, the lower temperature limit for denitrification (Paul and Clark 1989). The soil temperature began to increase in early March, reaching approximately 15°C. The nitrate concentration in the irrigation water ranged from 13 to 23 mg N L−1 (Fig. 2). The average concentration among the three plots and throughout the measurement period was 18 mg L−1. There were no significant differences in the nitrate concentrations of the irrigation water among the three plots. The average depth of floodwater in the three plots was 0.025 ± 0.014 m (mean ± standard deviation). The floodwater depths fluctuated as a result of precipitation events, but there was no significant difference among the three plots.
The nitrate concentration of the surface water at 2 days after insertion of the tubes decreased more in the RS plot than in the control or RSC plots (Fig. 2). In each treatment, large decreases in the nitrate concentration were observed just after irrigation (during December), but this decrease was not detected in January or early February, after which there was a marked decrease in mid-February in the RS plot. In contrast to the RS plot, only small nitrate decreases were observed in the control and RSC plots in this latter period (after mid-February). There was a significant difference in the surface water nitrate concentration between the control and RS plots (P < 0.05), but no difference between the control and RSC plots. These results show that no nitrate removal was obvious with any treatment during the lowest temperature period; however, incorporated RS became more effective in nitrate removal as the soil temperature increased.
The results of the denitrification potential measurements from the laboratory study supported the findings of the field study: the denitrification potential increased at low temperatures after the incorporation of RS (Fig. 3). Although the denitrification potential decreased with temperature in all three soils, soil from the RS plot showed significantly more denitrification potential than soil from the control at temperatures of 5 and 10°C. No significant difference was observed between the RSC and control plots. One reason for this could be the enrichment with C. The C sources in the RS plot were threefold higher than those in the RSC plot (Table 1). The amount of DOC extracted from the soils supports the enhancement of C in the RS plot. The DOC extracted from the control, RS and RSC plots was 0.12, 0.15 and 0.12 mg C g−1, respectively. Only the RS plot showed a 1% significant difference from the control. This enhancement of soil C should increase the microbial activity in the RS plot, as reported by Starr and Gillham (1993), who showed that sufficient labile organic C is necessary for denitrification. The important difference between our results and the results of Starr and Gillham (1993) is that our results show an increase in denitrification at low temperatures, whereas their results were obtained at higher temperatures (27°C). However, it appears that the quality of organic C also affects the denitrification potential. Labile fractions of RS, such as cellulose, protein and hemicellulose, decompose during the composting process (Chefetz et al. 1998). The rate of decrease of such C fractions would affect the denitrification potential. Because we did not determine the composition of the incorporated organic C, the effect of the quality of the C source remains unclear. We estimated the cumulative nitrate removal per unit area of paddy fields. Although these values vary with the residence time of the irrigated water (Tabuchi et al. 2001), our estimations were made under conditions in which the residence time was a constant 2 days, and were obtained from daily summation of each measurement. Comparing the calculations for each plot (Fig. 1), the cumulative nitrate removal from the RSC plot was 2.2-fold higher than the removal from either the control or the RSC plots. The control and RSC plots were not different in terms of cumulative nitrate removal.
The nitrate concentrations in the percolated water in the plots were quite low throughput the measurement period (Fig. 2). At a depth of 0.1 m, the nitrate concentration was lower in the RSC plot than in the other plots, which implies that incorporation of RS decreases the nitrate concentration not only in the surface water, but also in the percolated water at a depth of 0.1 m. The nitrate concentrations at 0.2 m were much lower than the concentrations at 0.1 m in the three plots. As noted by Nira and Atsumi (2007), efficient nitrate removal occurs in percolated water, even during low-temperature seasons, without the incorporation of organic material. Our results did not differ from theirs. The incorporation of RS does not significantly affect the net nitrate concentration of percolated water.
The results of the present study show that incorporation of RS into paddy fields is effective in increasing the amount of nitrate removal in winter to early spring. Our results are the first findings to suggest a suitable paddy management technique for nitrate removal in the winter season. A greater understanding of the best timing for the incorporation of organic matter, as well as the best types of organic matter for supplying adequate amounts of easily decomposable organic carbon, are required to fully develop this technique.
We thank Dr Kiyoshi Matsuo of the National Institute of Vegetable and Tea Science for helpful discussions and for arranging use of the experimental fields. We also thank Mr Hitoshi Sawase for the design and maintenance of the experiment fields. This study was funded through a plant breeding and environmental research project from the prefectural institutes of the Ministry of Agriculture, Forestry and Fisheries.