Reduction in data collection for determination of cultivar coefficients for breeding applications

Plant breeding applications of crop simulation models require cultivar coefficients of new breeding lines. These cultivar coefficients are normally estimated from field experiments conducted under optimum conditions over several environments with elaborate data collection for each line throughout its life cycle. Such an intensive sampling scheme poorly applies to breeding lines at the early testing stages, because the number of lines is large and seed supply is limited. The objective of this study was to determine the minimum data to be collected for the estimation of cultivar coefficients of peanut lines for breeding applications of the CSM-CROPGRO-Peanut model. Nine peanut lines varying in maturity were selected for this study. Data on plant growth and development stages of these lines that were collected following the recommended procedure were obtained from a previous study. These data were used in a stepwise procedure to determine the minimum plant characteristics required for deriving the cultivar coefficients. This included both the intensity of data collection as well as the type of phenological and growth characteristics. The full versus partial data sets were then used for model calibration to derive the cultivar coefficients and an independent data set was used for model evaluation. The results showed that (i) different types of reduced phenological data resulted in the same values of the cultivar coefficients; (ii) cultivar coefficients derived from some reduced growth data sets were as good as those derived from full data collection; (iii) model calibration of cultivar coefficients derived from various types of reduced data collection worked well for all development characteristics and fairly well for the normalized root mean square error (RMSEn) values of the plant growth characteristics; and (iv) model evaluation showed good agreements between observed and simulated values for all growth and development characteristics. These results showed that it is possible to reduce the data collection for cultivar coefficients determination. The minimum data suggested is to determine two developmental stages, i.e., first flowering (RI) and harvest maturity (R8), and three plant samplings for growth analysis, i.e., around the stages of full seed (R6), physiological maturity (R7) and harvest maturity (R8). (C) 2007 Elsevier Ltd. All rights reserved.