Estimation of parameters for some dominant maize (Zea mays L.) cultivars of Iran for using in APSIM mechanistic model

Investigating the impacts of agricultural management on crop productivity in specific environments is easily possible through crop simulation models. These computer-based tools, provide the ability of making some decisions regarding the use of agriculture resources and inputs. However, prior to use of these models, they need to be fully evaluated under different locations and years. Hence, this study has been performed to evaluate APSIM crop model and estimate the cultivar-specific parameters as inputs for the model.
Different datasets were used in order to parameterize and evaluate APSIM-Maize model for three maize cultivars. So that, a four-year dataset (Shiraz) was used for the calibration of K.SC 260 cultivar from early maturing group. Also, for K.SC 704 (from late maturing group) and Maxima (from mid-maturing group) were applied from the two regions (Kerman and Khoramabad, each one at one year). These datasets are used to adjust parameters based on biomass, grain yield, the number of days from emergence to flowering, the number of days from flowering to physiological maturity and LAI for each cultivar. Model validation was conducted by a series of other datasets including published articles in journals and final reports of research projects.
Three statistical indicators including coefficient of determination (R2) root mean squared error (RMSE) and index of agreement (d value) were computed from observed and simulated variables. The OriginPro 9.1 software was used to statistical analysis and draw figures.
The results of the model evaluation indicated that the model simulated flowering and physiological maturity stages of three cultivars with high accuracy in different years and different locations, so that the values of root mean square for flowering and maturity dates were as 1 and 1.35 %, respectively. Also, the simulation of LAI trend for two K.SC 704 and Maxima revealed that the APSIM-Maize model simulated LAI trend of these two cultivars with high accuracy. The values of root mean square for LAI trend of K.SC 704 in Kerman and Khoramabad were as 16.7 and 9.48 %, respectively. Also, for Maxima in Kerman and Khoramabad were as 16.75 and 10.65 %, respectively. Ultimately, APSIM-Maize model exactly simulated grain yield and biomass with high accuracy by accurate simulation of leaf area index and phenological stages.
Genetic parameters were different among the cultivars. The highest (650) and lowest (545) maximum number of kernel per ear belonged to cultivars K.SC 704 and K.SC 260, respectively. Also, the highest and lowest thermal time accumulations from flowering to maturity and those of seedling emergence to end of juvenile phase belonged to cultivars K.SC 704 and K.SC 260, respectively. Investigations show that the K.SC 260 is better rather than other cultivars just in terms of grain growth rate (9.6 mg kernel–1d–1). Finally, among the three cultivars, K.SC 704 has the highest biomass (21.81 t ha-1).
The evaluation of APSIM-Maize model indicated that the model simulated phenological stages (flowering and physiological maturity) with high accuracy which proves that this model has suitable structure for simulation of phenological stages. Also, among the different cultivars, the number of kernel per ear, thermal time accumulation from flowering to maturity and thermal time accumulation from seedling emergence to end of juvenile phase, were greater in K.SC 704 rather than other cultivars.