J.F. Angus1*, P.E. Bacon2, R. F. Reinke3*
1CSIRO Agriculture and Food, GPO Box 1600, Canberra, ACT 2601, Australia *corresponding author
2Woodlots and Wetlands Pty Ltd, 220 Purchase Rd, Cherrybroook NSW 2126, Australia
3International Rice Research Institute, Los Baños, Philippines
Nitrogen use efficiency (NUE) of flooded rice is notoriously low compared to upland crops. An exception is irrigated rice growing in the semi-arid, temperate Riverina region of south-eastern Australia where NUE is high and the time of fertiliser application is unusual by international standards. In this region, N fertiliser applied at the permanent flood (PF) stage leads to larger yield responses than when topdressed at panicle initiation (PI). In many other regions topdressing at PI is more efficient. In the field experiment described here, application of 200 kg N ha-1 as urea immediately before PF increased yield of medium-grain rice from 7.3 to 13.6 t ha-1, representing an apparent recovery of 76% of the applied N, compared with a yield of 11.4 t ha-1 and apparent recovery of 39% for same amount of N topdressed at PI. Sequential sampling showed that soil ammonium fell to low concentrations soon after urea application at both PF and PI and remained at background levels similar to the zero-N control for the rest of the growing season. Meanwhile above-ground N content of the crop increased steadily until maturity, suggesting that 38 kg N ha-1 of the fertiliser N had been temporarily immobilised before re-mineralisation and uptake by the crop. The higher efficiency of the PF application was because urea was washed into the soil with the irrigation water, while the urea topdressed at PI was initially held in the water column and at the soil-water interface before crop uptake, temporary immobilisation and loss, presumably via denitrification or ammonia volatilisation.