Abstract Sap flow sensors are crucial instruments to understand whole-tree water use. The lack of direct calibration of the available methods on large trees and the application of several data-processing procedures may jeopardize our understanding of water uptake dynamics by increasing the uncertainties around sensor-based estimates. We directly compared the heat ratio method (HRM) sap flow measurements to water uptake measured gravimetrically using the cut-tree method on a large mature aspen tree to quantify those uncertainties for ten consecutive days. The influence of the azimuthal position of the sensors and the application of different data-processing procedures on the accuracy of the sap flow sensors’ estimates was assessed using different metrics. Overall, the sap flow measurements showed high temporal precision with the gravimetric data. Azimuthal and radial variability of measured sap flux density showed the most substantial effect on the accuracy of the sensors’ estimates of whole tree sap flow. The zero-flow corrections applied altered the accuracy and linearity of the sensors’ measurements at the hourly scale, while the sapwood area method used had a lesser impact. Across the ensemble of available data-processing procedures, the cumulative whole-tree water uptake estimates for five consecutive days from the sensor diverged from the gravimetric measurements by less than 1% to more than 50% depending on the sensors azimuthal position and data corrections applied. This study illustrates some of the uncertainties associated with the methodological approaches chosen when using sap flow sensors to estimate water uptake in tall and large diameter trees.

中文翻译：

---

在大型成熟树上使用切割树方法对树液流测量数据处理程序引入的不确定性进行量化

摘要 树液流量传感器是了解整棵树用水情况的重要工具。缺乏对大型树木的可用方法的直接校准以及几种数据处理程序的应用，可能会增加基于传感器的估计的不确定性，从而危及我们对吸水动态的理解。我们直接将热比法 (HRM) 树液流测量与使用大型成熟白杨树上的切割树法测量的吸水量进行了直接比较，以连续十天量化这些不确定性。传感器的方位角位置的影响和不同数据处理程序的应用对液流传感器估计精度的影响使用不同的指标进行评估。总体而言，树液流测量显示重力数据具有较高的时间精度。测量的树液通量密度的方位角和径向变化显示出对传感器对整个树液流的估计准确性的最显着影响。应用的零流量修正改变了传感器在小时尺度上测量的准确性和线性度，而所使用的边材面积方法的影响较小。在可用数据处理程序的整体中，传感器连续五天的累积整棵树吸水量估计与重量测量值的偏差小于 1% 到超过 50%，具体取决于传感器的方位角位置和应用的数据校正. 这项研究说明了与使用树液流量传感器估计高大和大直径树木的吸水量时选择的方法学方法相关的一些不确定性。