ABSTRACTTraditional photosynthesis‐driven growth models have considerable uncertainties in predicting tree growth under changing climates, partially because sink activities are directly affected by the environment but not adequately addressed in growth modelling. Therefore, we developed a semi‐mechanistic model coupling stomatal optimality, temperature control of enzymatic activities and phenology of cambial growth. Parameterized using Bayesian inference and measured data on Picea abies and Pinus sylvestris in peatland and mineral soils in Finland, the coupled model simulates transpiration and assimilation rates and stem radial dimension (SRD) simultaneously at 30 min resolution. The results suggest that both the sink and phenological formulations with environmental effects are indispensable for capturing SRD dynamics across hourly to seasonal scales. Simulated using the model, growth was more sensitive than assimilation to temperature and soil water, suggesting carbon gain is not driving growth at the current temporal scale. Also, leaf‐specific production was occasionally positively correlated with growth duration but not with growth onset timing or annual cambial area increment. Thus, as it is hardly explained by carbon gain, phenology itself should be included in sink‐driven growth models of the trees in the boreal zone and possibly other environments where sink activities and photosynthesis are both restrained by harsh conditions.