Calibration and Validation of APSIM Millet Model for Proso millet (Panicum miliaceum L.) Accessions as a Basis for Crop Diversification

Purpose: Proso millet (Panicum miliaceum L.) is an underutilised minor millet grown as a rain-fed crop in subsistence farming systems in tropical African and Asian countries including Sri Lanka. It is identified as a climate-resilient crop that exhibits a huge potential to diversify conventional farming systems, however few modelling studies have assessed this crop, especially in the tropics. The objective of this study is to calibrate and validate Agricultural Production Systems Simulator (APSIM) model for Proso millet.Research Method: The millet model of APSIM that was developed based on Pearl millet was calibrated for five farmer-selected Proso millet accessions using the data gathered from two field experiments. Yield data collected from 35 farmers' fields under different crop management practices were used to validate the model.Findings: The observed phenology, leaf area index, above ground biomass and grain yield were not significantly different (p > 0.05) from the simulated values of the calibrated model. The observed yields (1187 +/- 336 kg ha-1) in the farmers'fields were not significantly (p > 0.05) different from the simulated yields (1208 +/- 255 kg ha-1), suggesting a good calibration of the model.Research Limitations: Lack of field data under different abiotic stress conditions is a limitation for further validation of the model.Originality/ Value: APSIM millet model has not been tested for Proso millet previously. Derived genetic coefficients were successfully used to simulate Proso millet production.

Automated summary

The purpose of this study was to calibrate and validate the APSIM model for Proso millet. The model was calibrated for five Proso millet accessions using data from two field experiments, and validated using yield data from 35 farmers' fields. The results showed that the simulated yields were not significantly different from the observed yields, indicating a good calibration of the model.