Evaluating cultivar intensity and dataset size for reliable cultivar recommendation in winter wheat: A systematic research of environmental and genotype factors

Abstract

Crop yield is influenced by environmental, genotype, and management factors. This study focuses on the environmental and genotype factors, specifically the concept of mega-environments, where similar crop varieties thrive due to similar environmental conditions, and cultivar intensity, a cultivar's favorable reaction to improved growing conditions, in cultivar recommendation for winter wheat in Poland. The research aims to evaluate the potential of using cultivar intensity as a tool for cultivar recommendation and investigate the influence of dataset size on model performance. The study utilizes a dataset of winter wheat grain yield data collected over six seasons from 19 experimental stations in Poland. Various models are compared using prediction measures, such as correlation coefficient, root mean square error, and mean absolute percentage error. The results show that models combining mixed analysis of variance and linear regression perform best in terms of yield prediction, followed by models using only regression. Models based on cultivar mean in the region exhibit lower prediction ability. The impact of dataset size on prediction accuracy is found to vary depending on the model and prediction measure used. The findings highlight the importance of considering dataset size when assessing model performance and emphasize the need for reliable data in cultivar recommendation. The outcomes of this study contribute to the understanding of cultivar recommendation strategies and provide insights into the use of cultivar intensity and dataset size optimization for reliable and accurate recommendations.