A comparison of cardinal temperatures estimated by linear and nonlinear models for germination and bulb growth of forage brassicas

Forage brassicas are widely used as a supplementary feed in temperate pasture systems but there is a lack of quantitative data about their growth and development. Furthermore, numerous models are available to estimate cardinal temperatures but there is uncertainty about whether linear or nonlinear models should be used. Initially a germination experiment was used to describe the rate response of nine forage brassicas to temperature. Three models were compared to estimate cardinal temperatures and the two best models were used for thermal time (Tt) accumulation for three groups of forage brassicas. Cardinal temperatures, defined as the base (Tb), optimum (Topt) and maximum (Tm), differed among groups of species for the bilinear and Lactin models but were similar within a group of species for these models. In most cases, cardinal temperatures estimated by the bilinear and Lactin models for the B. rapa group ranged from 3 to 4°C for Tb, 31°C for Topt and 41 to 48°C for Tm. For the B. napus and B. napus biennis groups these temperatures ranged from 0 to 3°C for Tb, 29 to 33°C for Topt and 38°C for Tm. The B. oleracea group had temperatures from 0 to 1°C for Tb, 25 to 27°C for Topt and 35°C for Tm. A second data set based on hypocotyl thickening was used to estimate the base temperature (Tb) for bulb growth of turnips and swedes. Both models estimated an average Tb of 4.2°C for bulb turnips and an average of 3.7°C for swedes. The Lactin model was considered the most adequate model to describe temperature responses where as, in some cases, the bilinear model had to be modified to account for changes in the rate of development. More importantly, an appropriate range of test temperatures was crucial for the estimation of reliable cardinal temperatures, independent of the model used.