An evaluation of irrigation-yield response models for peas (Pisum sativum L.)

This report deals with the evaluation of irrigation-yield response models for use with peas in Canterbury, New Zealand. The primary objective was to identify suitable models and to describe and predict the yield of peas due to irrigation and drought. The study uses seven crop seasons of experimental data [between 1973-1982], from the Winchmore Irrigation Research Station, Ashburton, Canterbury. The nine selected models may be conveniently classified into four groups: [1] Input-Output Models [2] Potential Deficit Models [3] Non-phasic Actual Deficit Models and [4] Phasic Actual Deficit Models. In these models the potential evapotranspiration was calculated using the Penman formula. Adjustment for crop cover and soil moisture availability were also included as required by certain models. Peas at maturity for dried seeds had a mean degree days of 1049 with a standard deviation of 102.60 and a mean photothermal time of 524 with a standard deviation of 46.10.1he mean crop period was 119 days during the experimental crop seasons. The total ‘water received’ model of the input-output group was least successful in describing the response of yield to irrigation. The other six models [number of irrigations, active ET, lost time, drought days, actual ET, Stewart S.1] from the input-output group, the potential deficit group, and the non-phasic actual deficit group] explained the pea yield variation due to irrigation and drought up to 50% and was found to be useful and appropriate for the purpose. The potential deficit models [active ET and lost time] suggested a limiting deficit of 34 to 36 mm [42.5% to 45% of available soil moisture] for providing irrigation to peas. The phasic actual deficit models [Jensen and Stewart S.2] described well the effect of different phases on yield response of peas, and incorporated many experimental evidence of the irrigation of peas into the model. Besides, it suggested further need for specific field experimentation for quantifying the effectiveness of different phases on pea-yield. The overall predictive ability of the models centred around the maximum pea yield obtainable in the field.