Efficiency of nitrogen fertilization of dry-seeded rice in south-east Australia

The efficiency and fate of fertiliser nitrogen applied to combine-sown rice were investigated in field experiments in which fertiliser timing, water regime and soil type were varied. The information gained was used to predict strategies for optimising fertiliser efficiency using current techniques, and to predict and test improved fertilisation technologies. The stage of crop growth, water regime and soil properties all had large and interacting effects on agronomic efficiency. On an infertile alkaline grey clay soil, agronomic efficiency of urea applied at sowing was very low (8 kg kg-1) compared with later applications associated with continuous flooding (up to 56 kg kg-1). The low efficiency was due to nitrification and subsequent de-nitrification during the flushing period. Nitrogen-15 balance studies indicated that 80% of the urea nitrogen was lost from the soil-plant system. The rapid nitrification rate and high loss of nitrogen on the grey soil contrasted with the very low levels of nitrate and low losses of nitrogen (10-25%) from urea applied to rice growing on an infertile acidic red soil under alternating conditions of saturation and aeration (sprinkler-irrigation). Furthermore, on a fertile acidic red soil in the same region, other authors have measured large yield responses to nitrogen applied at sowing. These differences highlight the need to consider soil properties and water management when attempting to predict optimum fertilisation strategies. Yields of sprinkler-irrigated rice (managed to replace water lost by evaporation) were reduced by more than 50% compared with rice grown under continuous flood. However, this was not due to decreased plant uptake or increased loss of fertiliser nitrogen in the sprinkler-irrigated treatments. The low yields appeared to be associated with plant water stress, decreased nitrogen mineralization (by approximately half) and location of the fertiliser nitrogen near the soil surface where root activity was probably restricted due to inadequate moisture. On the grey soil, maximum agronomic efficiencies occurred with application shortly before permanent flood or at early panicle elongation. Efficiency was doubled when urea was applied before permanent flood compared with shortly after permanent flood. The greater efficiency appeared to be associated with the deeper transport of the applied nitrogen into the soil, and consequently lower losses by ammonia volatilisation and/or nitrification/de-nitrification. However, even with the most efficient fertilisation strategies, plant 15N recoveries were less than 40%, while losses exceeded 20%. When the 15N balance data were considered in conjunction with the agronomic data,. it appeared that it would be possible to further increase agronomic efficiency if plant recovery of applied nitrogen could be increased. In particular, minimisation of losses of nitrogen (via nitrification/de-nitrification) from fertiliser applied before permanent flood was a most attractive option. Potential methods identified for increasing agronomic efficiency by minimising losses of nitrogen applied before permanent flood were deep placement and the use of nitrification inhibitors and slow release nitrogen sources. Several experiments were conducted in an attempt to improve fertiliser efficiency by banding urea and modified urea sources 5-7 cm below the soil surface before permanent flood. Plant recovery of 15N was increased by up to 20% with banding compared with surface broadcasting. The best recoveries were from urea super-granules (USG). An experimental fertiliser rig and a commercial seeder fitted with a triple disc assembly were used to band the fertilisers in the main plots. There was no significant yield advantage with banded urea over broadcast urea applied before permanent flood. The fertiliser rig caused considerable plant damage. Using the triple disc applicator, it was possible to band fertiliser below the soil surface with minimal soil disturbance, and plant damage, but only under i ideal conditions of soil moisture. Furthermore, where the soil surface was dry and cracked, penetration with the triple discs was no greater than the depth of the cracks into which surface applied urea prills would be washed upon flooding. The yield test of USG was unsatisfactory because of the excessive plant damage with this method. With current technology, there appears to be little scope for improving the efficiency of urea applied before permanent flood by mechanical placement below the soil surface