The application of N increased grain yields from 3.4 to 4.4 t ha−1. There was no interaction between N rates and water management, and no effect of water treatment on the grain yield. Before anthesis, dry matter accumulation increased with N application, but after anthesis, it decreased with increasing N application. With high N applications, grain filling was limited by a low contribution of post-anthesis assimilates. This indicates that traditional lowland rice fertilizer application schedules are not suitable for aerobic rice. N rate and timing of N fertilizer should be adjusted to balance crop's demand before and after anthesis. Reducing N before anthesis and applying N at anthesis is necessary to increase post-anthesis dry matter accumulation and finally grain yield. Furthermore, germplasm improvement is suggested to increase the translocation of dry matter to grain for increasing grain yield.
On average, yield was 14 % higher at the “experimental farm” than at the “experimental station”. Differences in crop growth and Mn uptake between the farm and the station mainly resulted from the difference in hydrology. In target areas for aerobic rice with a deep groundwater table more research is needed to determine whether Mn deficiency is a potential cause of yield suppression of aerobic rice.