Sap flux (F_t) measurements are extensively used to scale-up canopy transpiration and conductance, but time lag between sap flux and canopy transpiration is a problem. As canopy transpiration is nearly synchronous with micrometeorological drivers, better understanding of the lag relationships between F_t and micrometeorological drivers and soil water conditions would benefit the up-scaling of canopy transpiration from sap flux. Time series modeling at different spatial and temporal scales can identify and incorporate time lag effects, as well as multiple variables affecting transpiration and the interactions between them. SARIMAX and GARCH hybrid models were used to capture seasonality and autoregressive conditional heteroscedasticity effects. Two univariate hybrid models were designed to measure vapor pressure deficit (VPD_t) and photosynthetic active radiation (PAR_t), and one multivariate hybrid modelwas used in each season. Sap flow lagged behind canopy transpiration by 0–10 min in the dry season and 10–20 min in the wet season. VPD_t had a stronger influence on transpiration than PAR_t . Only the interaction between VPD_t and PAR_t in the wet season was observed. This study extends the application of time series modeling to the prediction of sap flow dynamics.