Evaluating a cassava crop growth model by optimizing genotype-specifc parameters using multienvironment trial...

Area of Science:

• Agricultural Science
• Plant Breeding
• Computational Biology

Background:

• Cassava (Manihot esculenta) is vital for sub-Saharan African food security.
• Crop growth models (CGMs) aid breeding by predicting performance across diverse environments and future climates.

Purpose of the Study:

• To assess the feasibility of large-scale CGM calibration within a cassava breeding program.
• To identify systematic biases in the CROPGRO-MANIHOT-Cassava model.

Main Methods:

• Parameterized the CROPGRO-MANIHOT-Cassava model using data from 67 clones across eight Nigerian locations (2017-2020).
• Employed trial-and-error adjustments and the General Likelihood Uncertainty Estimation (GLUE) method for calibration.
• Evaluated model performance using Pearson correlation, root mean squared error (RMSE), and d statistics.

Main Results:

• Post-calibration, correlation improved from -0.03 to +0.08, RMSE decreased from 21 t ha⁻¹ to 5 t ha⁻¹, and d increased from 0.23 to 0.44.
• The model underestimated root yield in dry, hot environments.
• The model overestimated root yield in wet, cool environments.

Conclusions:

• CGM calibration can be integrated into routine cassava breeding data analysis.
• Opportunities exist for refining the CROPGRO-MANIHOT-Cassava model to enhance prediction accuracy.