A plant growth model for integrated weed management in direct-seeded rice - I. Development and sensitivity analyses of monoculture growth

A new model, DSRICE1, was developed to analyze weed management strategies i n direct-seeded rice (Oryza sativa) systems. Previous rice models have not accounted for important cultural and weed management factors in direct-seed ed systems, such as growth from seeds and water-depth effects on plant grow th. Here we describe the development and sensitivity analysis of DSRICE1 fo r monoculture rice growth under water-seeded conditions. DSRICE1 is largely process-based and includes all standard weed management practices except f ertility. Simulation inputs include latitude, daily solar radiation, daily maximum and minimum temperatures, water depth. and seed rate. Phenology dep ends on thermal units. Growth begins with seed storage mobilization to seed lings, and photosynthesis starts when the first leaf is extended. Canopy li ght dynamics depend on leaf and stem area distributions for both live and d ead dry mass, and on water depth when submerged. Water-depth effects were e xplicitly simulated as reflection and attenuation of light. Model analyses revealed that parameter sensitivities varied over time. Some parameters wer e always important, while the effects of others were limited to particular parts of the season. Judged over the whole season, the most important param eters were for photosynthesis and light capture. Unlike in most monoculture simulations, early height gain rate was important in DSRICE1 because it de termined when plants emerged from the water into full light. Analyses of mo del structure and specifications revealed that predictions were significant ly affected by the use of skewed live leaf area distributions and the non-r ectangular hyperbola for the light response curve, and the inclusion of wat er-depth and dead canopy dry mass effects on canopy light dynamics. The cro pping system and management processes simulated in DSRICE1 had important ef fects on model predictions of rice growth. Explicit consideration of these factors distinguishes DSRICE1 from other rice growth models, and may lead t o better simulation analyses of system interactions with plant growth and w eed management strategies. (C) 1999 Elsevier Science B.V. All rights reserv ed.