Modelling the impact of drought-induced water stress on tree vitality and evapotranspiration

Droughts affect the vitality of trees and their central role in regulating ecosystems and the climate. Moreover, climate change is expected to increase the frequency and intensity of dry periods endangering the stability and functionality of forests. It is therefore essential to adequately represent drought impacts on forests in catchment scale eco-hydrologic models.

In this study, we aim at assessing the effects of water stress on tree vitality and evapotranspiration with the eco-hydrologic model SWAT+. To this end, the catchment of the Ellerbach (182 km² area) at gauge Schleifmühle and its tributary the Gräfenbach at gauge Argenschwang (32 km²) in the low mountain ranges of the Soonwald in southwest Germany were modelled. A spatially distributed parameterisation was applied to represent the spatial heterogeneity of the catchment. Model calibration was based on Latin Hypercube Sampling to derive 1000 parameter sets for eleven model parameters. From these model runs the best model run in terms of the smallest absolute percentage bias at both gauges was chosen. The model showed a good performance at the daily time scale at the catchment outlet and a lower but still acceptable performance in the upstream indicated by Kling-Gupta efficiencies of 0.81 (Ellerbach) and 0.66 (Gräfenbach) in the calibration period (2011 to 2016) and 0.83 (Ellerbach) and 0.64 (Gräfenbach) in the validation period (2017 to 2021).

Plant water stress was identified on a daily basis when the actual plant transpiration deviated from the potential plant transpiration. The highest water stress was found for forests in the low mountain range areas in the years 2011 and 2022, with durations ranging from 62 to 119 days in 2011 and from 43 to 130 days in 2022 for different tree species. These values are up to 2.1 times (2011) and 1.5 times (2022) higher than the long-term average. Generally, coniferous trees are more affected by water stress (long-term average: 91 days) than deciduous trees (long-term average: 31 days). However, in the two analyzed years deciduous trees experienced 2.1 (+34 days, 2011) and 1.5 (+15 days, 2022) times more water stress as compared to their long-term average, whereas coniferous trees experienced 1.3 (+25 days, 2011) and 1.4 (+34 days, 2022) times more water stress. 

Water stress affected tree vitality in the respective years indicated by the development of the leaf area index. Moreover, drought conditions led to a reduction in evapotranspiration by 14% (2011) and 12% (2022) when compared to mean annual evapotranspiration. Spatio-temporal differences in evapotranspiration patterns can be explained by an interplay between tree species, soil properties and precipitation patterns. As tree species perform different strategies for coping with water stress, future research shall focus on evaluating drought-induced transpiration decreases for the main tree species based on independent species-specific transpiration data.