PREDICTING PHASIC DEVELOPMENT OF GREEN BEANS FOR PROCESSING USING A MODEL WITH HIGH-TEMPERATURE REDUCTION OF THERMAL TIME ACCUMULATION

Green beans (Phaseolus vulgaris L. cv. Alcade) were sown in the field in Central Portugal in 1992 and 1993, on eight dates per year, under a ir temperatures, well above the expected optimum for plant development . It was assumed that the rate of development increases linearly with temperature from a base T-b to a sharply defined optimum T-o, beyond w hich this rate decreases linearly, reaching zero at T-c. An algorithm was constructed to fit this piecewise linear model (or any other model ) with daily or hourly temperature values. Cardinal temperatures of th e developmental process were, for example, 6.8, 24.0 and 34.1 degrees C, when hourly temperatures were used in the phase between emergence a nd the appearance of the first flower. The standard error (SE) of the estimates of the day of first flower was 2.73 days. A curvilinear mode l was also validated, but its performance was not significantly better . Using the fitted linear model, thermal time and SE of the estimates of the duration of the phases sowing-emergence, emergence-first trifol iate leaf, first trifoliate leaf-first flowering, first flowering-flow ering, flowering-pod development, and pod development-pod filling were calculated. The fitted linear model allows the computation of cardina l temperatures for development under high temperature conditions using limited field data, and can also be used to predict the occurrence of relevant phases of plant development. (C) 1997 Elsevier Science B.V.