In 1995, the denitrification profile through soil percolation of nitrate introduced by groundwater irrigation was studied in a rice paddy surrounded by a large area of upland fields. The denitrifying bacterial population was found to be higher in the suface layer of 0 to 2 cm than in the deeper plow layers during the entire submerged period from April to August. Denitrifying activity, as measured by the acetylene blockage method, was also found to be higher in the surface layer than in deeper layers at all times. Denitrification proceeded immediately after the start of incubation, with no supply of organic matter, suggesting that denitrification in the investigated paddy was not restricted by the depletion of organic matter during the submerged period and that denitrification can occur in the paddy as soon as nitrate is introduced by groundwater irrigation. In the interior of the paddy, nitrate in percolating water was nearly absent by 3 cm of depth and did not reach to the roots in the deeper plow layer throughout most of the submerged period. This is considered to be the reason the rice plants grew normally, with no damage caused by excess nitrogen uptake. In an exception to this, 90 d after submersion of the paddy, nitrate removal at the site near the water inlet was incomplete. Nevertheless, as this period is rather brief, the rice plants did not absorb excessive nitrogen and thus suffered no damage. In conclusion, the locus of denitrification is a notable factor in the actual application of nitrate removal in rice paddies, with the elimination of nitrate in the shallow plow layer being necessary.
