Observed and estimated leaf appearance of landrace and improved maize cultivars

ABSTRACT: The objective of this study was to compare the simulations of leaf appearance of landrace and improved maize cultivars using the CSM-CERES-Maize (linear) and the Wang and Engel models (nonlinear). The coefficients of the models were calibrated using a data set of total leaf number collected in the 11/04/2013 sowing date for the landrace varieties ‘Cinquentinha’ and ‘Bico de Ouro’ and the simple hybrid ‘AS 1573PRO’. For the ‘BRS Planalto’ variety, model coefficients were estimated with data from 12/13/2014 sowing date. Evaluation of the models was with independent data sets collected during the growing seasons of 2013/2014 (Experiment 1) and 2014/2015 (Experiment 2) in Santa Maria, RS, Brazil. Total number of leaves for both landrace and improved maize varieties was better estimated with the Wang and Engel model, with a root mean square error of 1.0 leaf, while estimations with the CSM-CERES-Maize model had a root mean square error of 1.5 leaf.

摘要：本研究旨在对比采用CSM-CERES-Maize（线性）模型与Wang和Engel（非线性）模型对地方品种及改良玉米品种出叶过程的模拟效果。本研究以2013年4月11日播种的地方品种‘Cinquentinha’、‘Bico de Ouro’及单交种‘AS 1573PRO’的总叶数数据集，校准两款模型的系数；针对‘BRS Planalto’品种，则采用2014年12月13日播种的数据集估算其模型系数。模型评估采用巴西南里奥格兰德州圣玛丽亚市2013/2014生长季（试验1）与2014/2015生长季（试验2）采集的独立数据集完成。相较于CSM-CERES-Maize模型，Wang和Engel模型对地方品种与改良玉米品种总叶数的预测精度更优，其均方根误差为1.0片叶，而CSM-CERES-Maize模型的均方根误差为1.5片叶。