Analyses of assumptions and errors in the calculation of stomatal conductance from sap flux measurements.

We analyzed assumptions and measurement errors in estimating canopy transpiration (E(L)) from sap flux (J(S)) measured with Granier-type sensors, and in calculating canopy stomatal conductance (G(S)) from E(L) and vapor pressure deficit (D). The study was performed in 12-year-old Pinus taeda L. stands with a wide range in leaf area index (L) and growth rate. No systematic differences in J(S) were found between the north and south sides of trees. However, J(S) in xylem between 20 and 40 mm from the cambium was 50 and 39% of J(S) in the outer 20-mm band of xylem in slow- and fast-growing trees, respectively. Sap flux measured in stems did not lag J(S) measured in branches, and time and frequency domain analyses of time series indicated that variability in J(S) in stems and branches is mostly explained by variation in D. Therefore, J(S) was used to estimate transpiration, after accounting for radial patterns. There was no difference between D and leaf-to-air vapor pressure gradient, and D did not have a vertical profile in stands of either low or high L suggesting a strong canopy-atmosphere coupling. Therefore, D estimated at one point in the canopy can be used to calculate G(S) in such stands. Given the uncertainties in J(S), relative humidity, and temperature measurements, to keep errors in G(S) estimates to less than 10%, estimates of G(S) should be limited to conditions in which D >/= 0.6 kPa.