Potential Added Value of Incorporating Human Water Use on the Simulation of Evapotranspiration and Precipitation in a Continental‐Scale Bedrock‐to‐Atmosphere Modeling System: A Validation Study Considering Observational Uncertainty

Abstract

Human activities, such as human water use, have been shown to directly influence terrestrial water fluxes and states. Simulations of soil moisture, river discharge, evapotranspiration, and groundwater storage are significantly improved, if human interactions, such as irrigation and groundwater abstraction, are incorporated. Yet improvements through the incorporation of human water use on the simulation of local and remote precipitation are rarely studied but may contribute to the skill of land surface fluxes. In this study, we evaluate the impact of human water use on the skill of evapotranspiration and precipitation in a fully coupled bedrock-to-atmosphere modeling platform. The results show that human water use can potentially increase the skill of the simulations across scales. However, observational uncertainty at the watershed scale limits the identification of model deficiencies and added value related to human water use. Locally, daily precipitation statistics potentially benefit from the incorporation of human water use. Although the incorporation of human water use does not remove the wet bias, it can increase the model skill.Plain Language Summary Precipitation forecasts exhibit large uncertainties, arising from unknown initial conditions and states, and limitations of models to represent processes at various scales. In order to improve simulations of the terrestrial water cycle, studies seek to improve the realism of the applied models through, for example, increased spatial resolution or the incorporation of additional processes across the Earth system. Here we analyze how the incorporation of additional processes in a European modeling system affects the accuracy of the simulated water fluxes, that is, evapotranspiration and precipitation. In particular, we improve the representation of groundwater in a continental-scale, atmospheric model and include groundwater pumping and irrigation, as major components of human water use. Results indicate that large-scale averages of precipitation, for example, over watersheds, are not necessarily improved if human water use is considered. However, significance of this finding is difficult to establish, because observational data sets exhibit large uncertainties, arising from, for example, the lack of observations in space and time, and miscalibration of the measuring devices. However, our results indicate that the accuracy of simulated daily precipitation is locally improved, which suggests that the incorporation of human water use may increase the accuracy of precipitation forecasts and advance our understanding of water cycle processes.

Using Long Short-Term Memory networks to connect water table depth anomalies to precipitation anomalies over Europe

“…Benefiting from the interaction of these components, LSTM networks show great promise for studying long-term relationships between time series. They have the ability to capture dependencies over 1000 time steps, outperforming standard RNNs whose upper boundary of reliable performance is only 10 time steps (Hochreiter and Schmidhuber, 1997;Kratzert et al, 2018). The response of wtd a to pr a is expected to exhibit a long time lag, especially in case of deep aquifers, and thus, LSTM networks are an appropriate type of network to use here.…”

“…These issues can be overcome by a variant of standard RNNs named Long Short-Term Memory (LSTM) networks (Supreetha et al, 2020). Although RNNs have been employed extensively in other science fields, particularly in natural language processing (D. , their application in hydrology is still in its infancy and has only recently received increasing attention (e.g., Kratzert et al, 2018;Shen, 2018;J. Zhang et al, 2018;Le et al, 2019;Sahoo et al, 2019).…”

Introduction of an Experimental Terrestrial Forecasting/Monitoring System at Regional to Continental Scales Based on the Terrestrial Systems Modeling Platform (v1.1.0)

“…TSMP, developed within the framework of the Transregional Collaborative Research Centre TR32 [16] has been applied in a wide variety of studies ranging from the regional to the continental scale with generally reasonable agreement to observations [17][18][19][20][21]. In addition, human water use has been implemented also assessing the potential added value in the comparison to precipitation and evapotranspiration products [22,23].…”

Continental-scale evaluation of a fully distributed coupled land surface and groundwater model ParFlow-CLM (v3.6.0) over Europe

“…spatial dynamics of water fluxes and state variables at higher resolution (12 to 1 km) over regional and continental scales (e.g., Keune et al, 2016Keune et al, , 2019Kollet et al, 2018;Tijerina et al, 2021;O'Neill et al, 2021). Despite these advancements, challenges still exist to implement and evaluate fully distributed integrated surface and groundwater models over large spatial domains, particularly given the lack of consistent large-scale hydrogeological information (de Graaf et al, 2020), and/or the computational cost to implement such models over larger domains.…”