Assessing different models to predict the growth and development of pepper plants under water deficits

Abstract

To construct pepper development simulation models under drought, experiments of water capacities of 45-55%, 55-65%, 65-75% or 75-85% and exposure (2, 4, 6 or 8 d) (Exp. 1 & 2), of 50-60%, 60-70% or 70-80% and exposure (3, 5, and 7 d) (Exp. 3) were conducted with "Sanying" pepper. Physiological development time (PDT), product of thermal effectiveness and PAR (photosynthetically active radiation) (TEP) and growing degree days (GDD) were used to simulate growth under various treatments in Exp. 1. Plant development was influenced by the severity and drought duration. Mild water deficits (65-75% for 2-6 d or 55-65% for 2-4 d) accelerated development, while severe water deficits (65-75% for 8 d, 55-65% for 6-8 d or 45-55% for 2-8 d) delayed development. The PDT gave the highest coefficient of determination (R², 0.89-0.94) and the lowest root mean squared error (RMSE, average of 1.03-1.50 d) and relative error (RE, average of 1.60-1.88%) for simulating three growth periods (Exp. 2). It was therefore used to construct growth models under water capacity of 45-85% over 2-8 d with spline, cubic, makima, linear, and nearest interpolation. Validation in Exp. 3 indicated that the spline model was optimal, having the highest R² (0.96-0.97) and the lowest RMSE (average of 1.31-1.75 d) and RE (average of 1.18-2.06%). The results of the study can help producers to optimize water management and to develop drought strategies for production.

Keywords: pepper; phenology simulation; physiological development time; spline interpolation; water capacity.