Abstract:Scientists have widely used sap flow measurements to estimate tree transpiration and to understand the physiological and environmental factors affecting the sap flow. However, the assumption of a uniform profile of sap flux may lead to a large bias when estimating whole-tree sap flow from measurements made with a single sensor in the outer part of the xylem sapwood. In this study, the sap flow velocities (Js) were measured at five depths (1 cm, 2 cm, 3 cm, 5 cm, and 8 cm) along a radial xylem transect of three larger trees of Populus alba L. var. pyramidalis Bunge, a popular tree species often planted in farmland shelterbelts of oasis in northern China. Measurements were made with five pairs of Granier thermal dissipation probes (TDP) for each tree over two consecutive growing seasons (in 2011 and 2012) to determine the radial sap flow pattern. The results showed that: (1) Js varies with depth in the descending order from 3 cm to 2 cm, 5 cm, 8 cm, and 1 cm. Js at depths of 1 cm (Js-1), 3 cm (Js-3), 5 cm (Js-5) and 8 cm (Js-8 ) were significantly and linearly correlated with that at 2 cm depth (Js-2), and their intercept coefficients were 0.24-0.27, 1.18-1.61, 0.81-1.64, and 0.38-0.75, respectively. The greatest Js appeared at depths of 3 to 5 cm under the cambium and the lowest appeared at depths of 1 cm or 8 cm; thus, the conventional 2-cm-long probe cannot reach the peak point of sap flow. Significant differences in Js were observed along the radial sapwood despite their similar diurnal course. (2) The atmospheric vapor pressure deficit (VPD) and solar radiation (Ra), two factors expressing the level of atmospheric vapor and heat, respectively, were the most important among the five meteorological factors controlling Js. The influence of Ra and VPD on Js in the outer and inner xylem during the growing season were similar for P. alba. However, accurately differentiating their respective contributions to sap flow proved to be difficult because of the interaction between them. The atmospheric potential evapotranspiration (ET0), which is a good indicator of the effects from meteorological variables, showed nearly similar diurnal pattern to Ra and was closely correlated with Js. (3) The diurnal course of Js for each depth displayed a near-normal curve with a peak at noon on typical sunny days. Js at depths of 2 cm to 8 cm peaked simultaneously. As an exception, the peak time of Js-1 appeared 4 min to 123 min ahead of the others. The peak times of all Js lagged behind Ra and were ahead of VPD. In July, P. alba exhibited vigorous physiological activities. The peak time of ET0, Js, and VPD on typical sunny days in July appeared at 12:30, 14:00 and 15:00, respectively. However, the time lag varied with season. The time lag between Js and ET0 (ΔJ-E) increased gradually from June to October with the amplitude of 70-110 min in 2011 and 70-128 min in 2012, indicating that solar radiation affected Js more strongly during the early growing season than later in the season. In contrast, the time lag between Js and VPD (ΔJ-V) decreased from June to October and varied in the range of 20-73 min in 2011 and 8-62 min in 2012, indicating that VPD affected Js more strongly in the late growing season than early in the season.
