Empirical models for harvest date prediction in broccoli (Brassica oleracea L. var. italica Plenck)

Broccoli is characterised by fluctuations in product supply. Therefore, harvest date prediction and programming continuity of supply is essential. Although crop growth and developmental models have been described in the literature, increased research on empirical models for harvest date prediction is needed. Here, data from 16 broccoli crops of 4 cultivars (Compacta, Comanche, Green Valient and Marathon), carried out all year round for two years, in three locations representative of the most important areas of broccoli production in Portugal, showed that a cold requirement was not detected for any of these cultivars and that, the time from planting to spear initiation could be described through a linear relationship between the reciprocal of time to spear initiation and the mean air temperature. From spear initiation to crop maturity the best empirical description was attained by using a quadratic relationship between thermal time, accumulated after spear initiation above 0 degreesC, and spear growth, although previous work to predict the duration of spear growth from initiation, for an early cultivar (Mercedes), was found to be closely related to an exponential model. From transplanting to harvest, the duration of crop growth can then be determined by combining the model for spear initiation with that for spear growth. From a field sampling of spear diameter and using long term average temperature data, the thermal time required for a spear to reach a certain diameter can be converted into the expected days to maturity. In addition, yield can be predicted by the linear relationship between spear diameter and spear fresh weight.