A model bridging waterlogging, stomatal behavior and water use in trees in drained peatland. (6th April 2022)

• Record Type: Journal Article
• Authors: Liu, Che; Wang, Qian; Mäkelä, Annikki; Hökkä, Hannu; Peltoniemi, Mikko; Hölttä, Teemu
• Editors: Diaz-Espejo, Antonio
• Is Part Of: Tree physiology. Volume 42:Number 9(2022)
• Page Start: 1736
• Page End: 1749
• Publication Date: 2022-04-06
• Subjects: hydraulic conductance -- Scots pine -- semi-process-based modeling -- soil flooding -- stomata; Trees -- Physiology -- Periodicals; 582.16
• DOI: 10.1093/treephys/tpac037
• Languages: English
• ISSNs: 0829-318X

Abstract

Abstract: Waterlogging causes hypoxic or anoxic conditions in soils, which lead to decreases in root and stomatal hydraulic conductance. Although these effects have been observed in a variety of plant species, they have not been quantified continuously over a range of water table depths (WTD) or soil water contents (SWC). To provide a quantitative theoretical framework for tackling this issue, we hypothesized similar mathematical descriptions of waterlogging and drought effects on whole-tree hydraulics and constructed a hierarchical model by connecting optimal stomata and soil-to-leaf hydraulic conductance models. In the model, the soil-to-root conductance is non-monotonic with WTD to reflect both the limitations by water under low SWC and by hypoxic effects associated with inhibited oxygen diffusion under high SWC. The model was parameterized using priors from literature and data collected over four growing seasons from Scots pine ( Pinus sylvestris L.) trees grown in a drained peatland in Finland. Two reference models (RMs) were compared with the new model, RM1 with no belowground hydraulics and RM2 with no waterlogging effects. The new model was more accurate than the RMs in predicting transpiration rate (fitted slope of measured against modeled transpiration rate = 0.991 vs 0.979 (RM1) and 0.984 (RM2), R 2 = 0.801 vs 0.665 (RM1) and 0.776 (RM2)). Particularly, RM2's overestimation of transpiration rate under shallow water table conditions (fitted slope = 0.908, R 2 = 0.697) was considerably reduced by the new model (fitted slope = 0.956, R 2 = 0.711). The limits and potential improvements of the model are discussed.