How and how accurately can we measure transpiration of forest stands? 

Water supply plays a crucial role in the vitality and growth of forests. Forests are an important source of high-quality drinking water and play a key role in the landscape water balance. Rising temperatures and changes in precipitation are leading to more frequent and severe droughts, which can increase mortality and pose a challenge for forest management and planning. At the same time, this is substantially changing the water balance of forests and thus the future seepage water supply. These challenges can be met with appropriate silvicultural measures such as tree species selection and thinning, but requires reliable knowledge about the transpiration of forest stands and the factors that influence it.The transpiration (evapotranspiration) of forest stands can be estimated using micrometeorological methods, e.g. eddy covariance measurements, or hydrological approaches, e.g. by measuring streamflow from clearly defined small catchment areas or large lysimeters with trees, or – due to their simplicity and availability often used nowadays – by measuring transpiration with sap flow sensors, with each method covering different scales and with each method including specific uncertainities. We developed a model system to improve the estimation of transpiration rates of forest stands based on sap flow and soil moisture measurements in combination with the stand water balance model LWF-Brook90 with the aim of integrating this approach into the Forest Environmental Monitoring Programme of ICP Forest https://www.icp-forests.net/. The combination of measurements and modelling aims at reducing the uncertainties and potential errors of estimates based purely on measurement data, as these alone contain a considerable degree of uncertainty in terms of their absolute values.Based on our 2-3 years measurements at three forest stands with simultaneous eddy covariance measurements as a reference, we would like to present the methodological approach and its improvements. Temporal dynamics of sap-flow measurements agreed well with EC data minus modelled evaporation, whereas absolute values were substantially over- or underestimated in most cases, indicating a high absolute uncertainty if estimates would be based on sap flow measurements alone. Comparing LWF Brook 90 modelling results of a set of established reference parameters to sap flow measurement dynamics showed that tree eco-hydrological parameters needed to be modified to yield a better agreement of modelled transpiration with the temporal sap flow dynamics, which is the more reliable information in the sap flow measurement. Together with the results from a further 13 forest stands in Germany the characteristic differences in the annual transpiration pattern and water consumption of theses stands will be presented.