Modeling the Development of Cultivated Rice and Weedy Red Rice

Abstract

Development models are an important part of crop simulation models and are useful tools for field operations. The objective of this study was to simulate the development of cultivated rice and weedy red rice with a linear model (thermal time model) and with a non-linear model (WE model). Data from a four-year experiment conducted during the 2003-2004, 2004-2005, 2005-2006, and 2006-2007 growing seasons in Santa Maria, Rio Grande do Sul State, Brazil, with nine cultivated rice genotypes and two weedy red rice biotypes were used. Plants were grown in 12 L pots during the four years and in a paddy rice field during the 2006-2007 growing season. Developmental data available and used in this study were dates of emergence (EM), panicle differentiation (R1), anthesis (R4), and all grains with brown hulls (R9) recorded in five plants per replication in each sowing date. Data collected in the 2004-2005 and 2005-2006 growing seasons were used to estimate the coefficients of the two models, and data collected in the 2003-2004 and 2006-2007 growing seasons were used as an independent data set for models evaluation. The root mean square error (RMSE) for all developmental stages varied from 4.9 to 10.5 days with the thermal time model and from 4.3 to 10.9 days with the WE model. The WE model gave better predictions in six out of eleven genotypes, with better predictions for early (R1) than for later (R4 and R9) developmental stages. The WE model described properly the development of rice plants grown in pots and in a paddy rice field and can be recommended as an alternative model to the thermal time model for rice in situations when the air temperature falls into the non-linear range of the response of development to temperature.