Two simplified methods for assessing the emission of ammonia, from fertilized, flooded rice (Oryza sativa L. cv. Inga) fields without affecting the plant's environment, are described and compared with an established micrometeorological method. Method A requires a circular plot of a 25‐m radius over which wind speed and atmospheric ammonia concentrations are observed at 0.8 m above the floodwater. Method B requires measurements of ammoniacal nitrogen concentration, pH and temperature in the floodwater, and wind speed at a fixed height. Method A is the more accurate technique and requires fewer measurements. It appears that one calibration curve relating vertical flux of ammonia to these measurements can be applied in widely different environments. Method B is not only less accurate than Method A but appears to require different calibration curves to determine vertical flux for different environments. However, this method may be more useful for comparing ammonia losses from several different fertilizer treatments since only small plot areas are required for each treatment. Method B can be calibrated easily for each site by one set of simultaneous measurements on one treatment plot, using Method A.

Direct measurements of gaseous loss, 15N labelling and intensive sampling of floodwater and soil columns were used to follow the transformations and movement of fertilizer nitrogen (N) applied as urea (80 kg N ha-1) into the floodwater of a young rice crop. During the first 11 days after urea application, about 46% of the applied N was lost from the water-soil-plant system. Only 11% was volatilized as ammonia, despite very high floodwater pH values (up to pH 10) and some strong winds. Gaseous loss as nitrous oxide and leaching of N beyond a soil depth of 100 mm were both negligible. All the evidence indicates that the other 35% of the applied N was lost as dinitrogen produced by denitrification of nitrite and nitrate after nitrification of ammonium near the soil surface. No further losses from the system were detected after the first 11 days. Final recovery of the fertilizer N by the above-ground crop was only 17%. The results show that the several processes which resulted in gaseous loss of N operated concurrently, commencing almost immediately after fertilizer application. The major microbiological transformations appeared to occur near the soil-water interface (c. 0-20 mm depth).

Emissions of nitric oxide and other odd nitrogen oxides (NOv) from a flooded rice field were studied after urea had been broadcast into the floodwater.The NO x flux from the fertilized area was very low (0.2 x l0 -9 g N m -2 s -I ) for the first few days after application of urea and was high (0.95 × 10 -9 g N m -2 S -1 ) in the subsequent period when significant nitrite and nitrate were present in the floodwater. At night, little if any NOx was exhaled but ambient NO2 was absorbed by the floodwater. An uptake velocity for NO2 of 3 x 10 -4 m S -1 was measured during one night. Maximum NOv losses were observed near 1300 h when temperature and solar ultraviolet light were maximum.While the amounts of nitrogen oxides emitted are of little agronomic importance (-2 x 10 -3 per cent of the fertilizer nitrogen was lost as NOv during the 10-day study period), they may well be of significance as a source for some gas reactions in the atmosphere and for the global nitrogen cycle.Of the fertilizer nitrogen applied (as urea) approximately 30% was lost to the atmosphere by NH 3 volatilization, 15% by denitrification, presumably as N 2, and the remainder, less minor losses of NO and N20, remained in the plant/soil/water system.

Long periods of waterlogging during irrigation are thought to adversely affect wheat growth and yield. This work quantifies these effects for wheat (Triticum aestivum L. cv. Bindawarra) growing on a transitional red-brown earth. The crop was irrigated on four occasions during spring. At each irrigation six ponding treatments were imposed involving sprinkler irrigation (SP) or flooding for 1, 12, 24, 48 or 96 h. Each plot always received the same treatment. Nitrogen fertilizer was applied at sowing (23 kg N ha-1 as ammonium phosphates) and at the end of tillering, 3 weeks before the first irrigation (100 kg N ha-1 as urea). There was also a 1 h ponding treatment which did not receive urea at the end of tillering. Grain yield on the 48 h and 96 h ponding treatments was significantly lower than on SP, 1 h and 12 h. Yield declined by 69 kg ha-1 for each day that water was ponded on the surface, or by 55 kg ha-1 for each day that the estimated soil water depletion was below 25 mm. The wheat crop suffered stress due to lack of aeration for periods varying from 42% (1 h and SP) to 68% (96 h) of the time interval between early stem elongation and physiological maturity. The grain yield and N uptake data indicate that the site was highly responsive to N fertilizer. However, there was no significant effect of ponding treatment on total N uptake, suggesting that the yield differences between the ponding treatments were not caused by differences in N availability. The results demonstrate that yield loss through waterlogging can be minimized by changing to irrigation management practices that minimise the period of ponding.