Modelling the effect of groundwater depth on yield-increasing interventions in rainfed lowland rice in Central Java, Indonesia

Because of drought and nutrient stress, the yields of rainfed lowland rice in Central Java, Indonesia, are generally low and unstable. Variation in groundwater depth can contribute to experimental variability in results of yield-increasing interventions. To test this hypothesis, we used the crop growth simulation model ORYZA2000 to explore the impacts of groundwater depth on the effect of sowing date, tillage, fertiliser-N application and supplementary irrigation on the yield of lowland rice at Jakenan, Central Java, Indonesia. ORYZA2000 was first parameterized and evaluated using data from eight seasons of field experiments between 1995 and 2000. The model adequately simulated the soil water balance, crop growth and grain yield. With shallow to medium groundwater depth (less than 0.5 m deep), rainfed rice yields are close to potential yields with timely sowing in the wet season. With groundwater tables fluctuating mostly between 0.5 and 1.5 m, rainfed yields are 0.5–1 Mg ha−1 lower than potential yields with timely sowing. The decrease in yield with late sowing sets in earlier and proceeds faster with deeper groundwater depths. Deep tillage and supplementary irrigation increase yield more with deep groundwater tables than with shallow groundwater tables, but N fertilisation increases yield more with shallow than with deep groundwater tables. Groundwater depth should be taken into account in the selection of yield-increasing interventions.