Modelling for maize irrigation scheduling using long term experimental data from Plovdiv region, Bulgaria

Independent historic datasets on irrigated maize, collected over seven years (1984-1990), were used to parameterize the irrigation scheduling model ISAREG. Experimental data were obtained under rainfed, deficit, and full irrigation conditions in an alluvial soil at Tsalapitsa, Plovdiv region, in the Thracian plain, Bulgaria. Crop coefficients and depletion fractions for no-stress were calibrated by minimizing the differences between observed and simulated soil water content. The calibration was performed using data from full irrigation and rainfed treatments while deficit irrigation treatments were used for validation. The modelling efficiency was high, 0.91 for the calibration and 0.89 for the validation. The resulting average absolute errors of the estimate for the soil water content were smaller than 0.01 cm3 cm-3. The model was also tested by comparing computed versus observed seasonal evapotranspiration. Results for dry years show a modelling efficiency of 0.96 but the model slightly underestimated evapotranspiration for other years. The yield response factor was derived from observed yield data of the hybrid variety H708 when relative evapotranspiration deficits were smaller than 0.5. The value Ky = 1.32 was obtained. The relative yield decreases predicted with this Ky value compared well with observed data. Results support the use of the ISAREG model for developing water saving irrigation schedules for the Thracian plain.