Near stream groundwater table fluctuations impact transpiration rates of riparian plants: a field study with stomatal conductance and dendrometry measurements

Vegetation establishment, growth, and succession in riparian ecosystems are linked to river and groundwater dynamics. This is especially true in Alpine gravel-bed rivers with wide floodplains and a strong river-aquifer exchange. Here we provide data evidence of riparian plant response to short-term groundwater table fluctuations in a braided gravel-bed river (Maggia). We used indirect physiological variables for photosynthesis and transpiration – stomatal conductance g_s and daily variation in stem diameter ΔD_d – which we measured at six mature riparian trees of the Salicaceae family, one Populus nigra and one Alnus incana specimen at two sites during two growing seasons. The site where g_s measurements were conducted showed a greater depth to groundwater with higher variability compared to the site were dendrometers were placed.

We analysed the data by means of two different random forest regression algorithms for the two study sites. One with the transpiration-induced daily tree diameter drop during the growing season 2017 as the dependent variable, and one with the raw g_s measurement sequence, obtained on 10 days throughout the growing season 2019, as the dependent variable. In both algorithms the independent variables consisted of meteorological measures (locally measured and at valley scale) and of groundwater and river stages near the individual plants. We also separated the g_s measurements into low and high groundwater stage conditions observed during the g_s field campaign and applied traditional regression analysis of g_s on vapor pressure deficit VPD and global radiation r_g for the 2 groundwater stage conditions separately.

The data analyses demonstrate that:

(a) short-term variation of the groundwater table affects riparian vegetation: at the site with deeper groundwater, the water table depth was the best predictor of g_s variability, while at the site with shallower groundwater, temperature and vapor pressure deficit were the best predictors of ΔD_d  variability;

(b) instantaneous stomatal conductance is related to vapor pressure deficit (VPD), but conditioned by groundwater levels, with higher stomatal conductance for the same radiative input and VPD when the water table was higher.

(c) local micro-climate measured at tree locations had a stronger predictive power for g_s than valley scale climate, suggesting local climate may be an important control on vegetated stands on gravel bars.

Even though the considered plants are located in close proximity to the river and could be considered to be unaffected by water stress, our analysis provides evidence of riparian trees undertaking physiological adjustments to transpiration in response to groundwater stage, depending on their riparian floodplain settings. In the heavily regulated Maggia river this has implications on the minimum flow release by dams, as prolonged periods of low water stage in the river will lead to a decrease in groundwater stage, and subsequently in reduced growth of phreatophytic riparian plants on the floodplain. We argue such plant-scale measurements should be helpful for the optimisation of flow release levels in regulated riparian systems.