Construction and primary evaluation of a simple model to predict potential dry matter production in pearl millet

A simple mechanistic model was constructed to evaluate potential dry matter of pearl millet. All needed parameters were calculated based on many field and laboratory experiments and actual data obtained from 86 fields of Millet located at different regions of Khorasan's provinces. The major simulated processes were crop phenology (as a function of temperature and photoperiod function curves), development of leaves (as a function of daily biomass partitioning to leaves and Specific Leaf Area (SLA)), and total dry matter (as a function of intercepted radiation and radiation use efficiency).The model was run based on daily time step and outputs were evaluated by comprising actual and predicted data. Calculated Root Mean Square Error (RMSE) by actual and predicted data indicated that predicted days to flowering biased 2.61 and 3.37 days from mean of actual values for the first and second sowing dates in studied regions, respectively. Determination coefficients of actual data plotted against predicted data revealed that predicted dry matter yield was acceptable in comparison with actual data. Although, the model had an overestimation during two-third of growing season, but low RMSEs (0.6 and 0.74 ton ha-1 for first and last sowing dates, respectively) demonstrated the ability of model to predict dry matter production during growing season of pearl millet (R2=0.98). In this study, the advantages and deficiencies of model, also, have considered.