Vegetation activity plays a key role in the water cycle, with transpiration accounting for 50–80\% of total terrestrial evaporation globally. Yet many conceptual hydrological models neglect phenology, estimating transpiration as a function of relative soil moisture and assuming constant vegetation activity. This simplification leads to systematic errors, particularly in winter, when conventional models often overestimate transpiration.

Because transpiration is strongly correlated with sap flow, this study developed three methods to incorporate tree phenology into a semi-distributed conceptual hydrological model. The model structure distinguished between coniferous and deciduous trees. Sap flow dynamics were included either directly, using sap flow data, or indirectly, using temperature as a proxy for seasonal variation.

To address the limited availability of sap flow data, a Generalized Additive Model (GAM) was developed to predict normalised sap flow. Using temperature, relative humidity, incoming shortwave radiation, volumetric soil water content, and normalised accumulated growing degree-day as predictors, the model reliably reproduced sap flow dynamics for both tree types. In addition, the GAM framework enabled separate analysis of each predictor's relationship with sap flow, providing clearer insights into the underlying processes and the relative influence of individual variables.

The results show that including tree phenology improves the ability of conceptual hydrological models to represent transpiration seasonality and vegetation dynamics. Although discharge simulations were not substantially improved, the added internal realism reduces equifinality and makes the model more robust under changing conditions.

Among the three developed methods, the direct inclusion of sap flow dynamics resulted in the highest model performance, particularly in transpiration simulations. This highlights the added value of the sap flow prediction model itself, which provided a robust link between environmental drivers and vegetation dynamics, thereby strengthening the integration of phenology into conceptual hydrological modelling.