Effects of temperature on head growth of broccoli (Brassica oleracea L. var. italica):: Parameter estimates for a predictive model

Scheduling harvests in broccoli crops by planting at time intervals may cause problems because of the temperature dependent development of the plant. Data from three years of field experiments with four plantings per year, three cultivars and two plant densities were used to estimate the effect of environment on broccoli growth. A quadratic relationship between the logarithm of head diameter and a temperature sum with base 0 degrees C and ceiling 17 degrees C starting at head initiation (diameter of apex 0.6 mm) accounted for 97.3% of the variation in head diameter for pooled data of 68 crops. Including effects of plant density and cultivar significantly improved the fit of the quadratic model (R-2:0.9819). Including accumulated solar radiation from head initiation or planting month as additional explanatory variates could further improve the fit of a quadratic relationship and to the same extent (R-2:0.9843% and 0.9847%). The simple quadratic relationship between broccoli head diameter and temperature sum from head initiation can be used in a predictive model to forecast when broccoli heads reach a specific size. The parameter estimates of a predictive model for head growth based on temperature sum is given for three cultivars at two densities. Cross validation of the model showed that 95% of the predictions of the temperature sum to a certain head size were in the range of 4-5 d from observed. The broccoli head was never initiated before at least seven visible leaves or a total of 14 leaves were initiated. A linear relationship between the number of visible leaves and the total number of initiated leaves explained 95% of the variation in data. Leaf number may be used to indicate the earliest possible stage to take samples for prediction of harvest time, but not to indicate the size of the head because of the temperature dependent induction.