Testing CERES-Sorghum model in simulating yield and yields component of forage sorghum in relation to plant population.

摘要

This study was conducted during kharif season of 1993 to compare forage biomass, yield and leaf area index (LAI) simulated with CERES-Sorghum model with actual biomass, yield and LAI produced across a wide range of plant population in Jhansi, Uttar Pradesh, India. The treatments consisted seven-plant population density: 20, 30, 40, 50, 60, 70, 80 row to row distance (cm) with 20, 15, 10, 6, 4, 3, 2 plant (m2). The experiment was laid out under optimal conditions (90 kg N/ha and 40 kg P2O5/ha) with three replications having randomized block design. The highest dry matter production (15050 kg/ha) and maximum LAI (5.01) were recorded under 15x104 plant/ha. Simulated LAI and total dry biomass (TDM) followed the same pattern throughout the growing period in high plant density, i.e. 10 to 20x104plants/ha. LAI and TDM could be predicted fairly well in treatments having high plant population without bias. In low plant population (i.e. between 6 to 2x104 plants/ha), the model tended to underestimate LAI and TDM. The mean predicted and observed TDM was 5234 and 11038 kg/ha, respectively, with mean difference (standard deviation) of -6103.0 (+556) kg/ha. However, the model generally over predicted the grain yield (bias=983.5 kg/ha) and 1000 seed weight (bias=13.9 g). Weakness/problem in the CERES-Sorghum model to underestimate TDM under condition of population below 10x104 plants/ha, and inability to partition dry matter into stover and grain yield high population must be resolved before using CERES-Sorghum model to address management decisions for forage sorghum..