Testing the performance of ORYZA1, an explanatory model for rice growth simulation, for Mediterranean conditions

ORYZA1 is an explanatory model to simulate rice growth, development and leaf area index (LAI) under potential production. The present study aims at testing the performance of ORYZA1 for Mediterranean conditions (farming practices, cultivars, weather) for fully irrigated direct-seeded rice. ORYZA1 was calibrated and validated with field data of two cultivars, a short-grain (Tebre) and a long-grain cultivar (L-202), grown in various years in the Ebro Delta of Spain. Phenological development of the rice crop, daily dry matter production and leaf area development were calibrated. Tebre and L-202 had no significant differences in the total length of the development period. The pre-heading period, however, was longer and the post-heading period shorter in L-202 than in Tebre. This induced differences in translocation characteristics, spikelet number per unit area, weight of the grains and harvest index. The following crop characteristics were similar between cultivars: extinction coefficient (increased with development stage), dynamics of nitrogen distribution, partitioning of assimilates, relative death rate of leaves, relative growth rate of leaf area during exponential growth, specific leaf area and a strongly decreasing specific stem green area. The simulated curve fitted much better the observations, which was clear from a strongly reduced value of RMSE, when considering that LAI comprises the leaf blade area only, without a photosynthetic contribution by stem green area. The model simulated rice growth very accurately until flowering. After flowering, however, divergences appeared and increased especially at the yellow ripe stage. From then on the crop did not grow much more, whereas it continued in the simulation. This reduction of growth rate was usually accompanied by an increase in the relative death rate of leaves and the drying of the grains. The main source of error may be a limited understanding of the ripening and sink limitation processes. A considerable yield gap between potential and observed yield remained. A climatic variability assessment over 10 years, from 1987 to 1996, showed a small but correlated variation (r=0.7) in both simulated and measured rice yields.