A simulation model for the dynamics of rice growth and development: III. Validation of the model with high-yielding varieties

Abstract

A wide range of applicability is a desirable feature of any simulation model. A crop model should therefore prove its validity for different varieties and various climatic conditions. Here, a rice ( Oryza sativa L.) crop model originally developed for Makalioka 34, a local variety grown in Madagascar, is validated with experimental data on growth and development of IR58 and IR64, two high-yielding varieties from the Philippines. The model simulates carbon and nitrogen dynamics and is based on demographic theory and the metabolic pool approach which has been successfully applied in population models across different trophic levels. 1. (1) The model results are compared with field data on dry mass growth of IR58 and IR64 at two locations, in three years and two planting densities. 2. (2) The simulation of nitrogen uptake and relative N content is validated with corresponding field measurements on IR64. 3. (3) Aside from visual comparison, the computation of a standardized bias and a standardized mean square error are proposed to quantify the goodness of fit. 4. (4) By means of simulation the sensitivity of Makalioka 34, IR58 and IR64 to nitrogen fertilization and temperature is evaluated. The model appears to have a sufficient validity range to produce reliable predictions on dry mass growth as well as nitrogen dynamics for different varieties and under different climatic conditions. It yields slightly better results for IR64 than for IR58 when comparing the standardized bias and standardized mean square error. Field data suggest that the model accurately simulates the effects of nitrogen fertilization in Makalioka 34 but underestimates the relative yield response to N applications in IR58 and IR64. According to the model, IR58 and IR64 seem to be well adapted to the actual temperatures in the Philippines while Makalioka 34 thrives better under lower temperatures. In its present form, the model proves to be a useful tool to investigate and explain the most important processes of yield formation for different rice varieties grown under various agronomic and climatic conditions.