Comparisons of three scaling up methods to estimate stand transpiration of a xerophytic shrub (Salix psammophila) in northern China

Abstract

Transpiration (T) is a key hydrological process, delivering water essential for plant metabolism and thus affecting productivity. A major challenge in estimating stand T is how to accurately scale sap flow data from individual trees to the stand. In shrub ecosystems, various scaling up methods have been used to extrapolate tree-level sap flow measurements to stand-level T, these include leaf area, cross-sectional area and number of branches. However, the performances of different scaling up methods have not been fully explored for shrubs. In this study, we measured sap flow of a xerophytic shrub (Salix psammophila) and scaled up using measures of leaf area, cross-sectional area and numbers of branches in order to estimate stand T during the rainy seasons in 2019 and 2020 on the northern Loess Plateau, China. In addition, we measured precipitation, throughfall, stemflow, soil evaporation, surface runoff and 0–200 cm soil water content to calculate actual stand T on the basis of soil water balance method. The results revealed that daily stand T differed according to the scaling up methods used for the estimation. Daily estimated stand T based on measures of leaf area (0.1–13.1 mm d−1) was consistently higher than those based on cross-sectional area (0.2–8.6 mm d−1) and number of branches (0.4–8.9 mm d−1) (p < 0.05). During the two rainy seasons, the actual daily mean stand T (2.5 ± 1.6 mm d−1) was significantly lower than the estimation by the three scaling up methods (p < 0.05). The method based on cross-sectional area appeared to be most suitable for scaling up because it had the lowest root mean square error and bias values (0.939 mm d−1 and 0.633 mm d−1, respectively). This study highlights the wide variations of stand T upon which scaling up method was chosen, and these differences need to be considered when converting tree-level sap flow to stand-level T in shrub and other ecosystems.