Half-hourly mean values of transpiration measured by eddy covariance over the course of six growing seasons in two boreal forest sites were used to develop stand-level relationships between transpiration and soil water content. The two sites were an aspen site on fine-textured soil and over five growing seasons for a jack pine site on coarse-textured soil in Saskatchewan, Canada. About half of the data record covered a multi-year drought that was more severe at the aspen site than the jack pine site. Measurements of transpiration and environmental variables were used to adjust a transpiration model to each site, with environmental variables retained in the model based on their capacity to improve the model adjustment. The model was also used to produce estimates of maximum canopy conductance (gcMAX). The fit of the model to the aspen half-hourly transpiration is better than to the jack pine data (r2 of 0.86 versus 0.60). Relative soil water content explains more of the variability in half-hourly transpiration at the aspen site (46%) than at the jack pine site (10%). The relationships between transpiration and environmental variables are stable throughout the drought suggesting an absence of acclimation. Published soil water modifier curves for loamy clay soils compare well with the modifier function we obtained for a similar soil at the aspen site, but the agreement between the published curve and our curve is poor for the sandy soil of the jack pine site. Values of gcMAX computed at the half-hourly scale are greater at the aspen site (14.3mms−1) than at the jack pine site (10.2mms−1), but we hypothesize that the coarse soil and perennially lower water content of the jack pine site may cause this difference. Finally, we also present values of gcMAX computed at the daily and monthly scales for use in models that operate at these time steps.