CERES model application for increasing preparedness to climate variability in agricultural planning – calibration and validation test

Abstract

The procedure of stepwise calibration and validation of CERES-maize and CERES-wheat models was used for models adjustment in two fields under contrastive soil conditions (Chromic Luvisol and Vertisol) in Sofia region. Both models reflected well the phenomenon of water/nitrogen extraction and retention in the root zone after the calibration. Adjusted CERES wheat was validated over a range of soils, varieties, climate, and management conditions and proved acceptable reliability of model predictions in most of the tested situations. The highest degrees of association ( R 2 > 0.85%) was established between observed and simulated series of potentially extractable soil water (PESW) on fertilised Vertisol and Chromic Luvisol when the error (RMSE) ranged from 8.9 to13.3 mm in moderately wet, moderate and dry wheat vegetation seasons. Detailed observations of the vulnerable “Chromic Luvisol-maize” agroecosystem in lysimeters, including water and nitrogen fluxes at the bottom boundary, enabled to validate integrally the prediction capacity of calibrated CERES-maize model. Simulation outputs after calibration about PESW, water uptake, drainage, soil nitrogen storage, nitrogen leaching and crop dry weights, proved to be accurate enough over a comparatively long period (1.05.1997–1.10.1999). Graphical and statistical test of CERES-maize output characterised the performance of the model as acceptable over the specific conditions of validation lysimeter trial.