Fully distributed hydrological models take into account the spatial variability of a catchment, and allow for assessing its hydrological response at virtually any location. However, these models can be time-consuming when it comes to model runtime and calibration, especially for large-scale catchments. Meanwhile, deep learning models have shown great potential in the field of hydrological modelling, but a multivariable, fully distributed hydrological deep learning model is still lacking. To address the aforementioned challenges associated with fully distributed models and deep learning models, we explore the possibility of developing a fully distributed multivariable deep learning model by using Graph Neural Networks (GNN), an extension of deep learning methods to non-Euclidean topologies. We develop a surrogate model of wflow_sbm, a fully distributed, physics-based hydrological model, by exploiting the similarities between its underlying functioning and GNNs. The GNN model uses the same input as wflow_sbm: gridded static parameters based on physical characteristics of the catchment and gridded dynamic meteorological forcings. The GNN model is trained to approximate wflow_sbm outputs, consisting of multiple gridded hydrological variables such as streamflow, actual evapotranspiration, subsurface flow, saturated and unsaturated groundwater storage, snow storage, and runoff. Our results show that the GNN model accurately predicts multiple hydrological variables in unseen catchments (median KGE=0.76), and can serve as an emulator of wflow_sbm with a shorter runtime. We furthermore demonstrate how the GNN model can function up to a prediction horizon of a full year, using physical system states to account for system memory, as well as a curriculum learning strategy combined with a multi-step ahead loss function during training. Overall, this study contributes to the field of fully distributed modelling using a deep learning approach.

The study of water is without question an important one. We can learn a lot by studying the theory but in the end, it is essential to actually go out into the field and see the hydrological processes in practice for ourselves. The educational value of fieldwork lies not only in seeing theory become reality but also in experiencing the difficulties and limitations of gathering data first-hand.

The water resources department of the Hanoi University of Natural Resources and Environment (HUNRE) wanted to expand its curriculum with a fieldwork excursion that will be part of three courses related to surface water, groundwater, and water quality. During our ten weeks in Vietnam, we assisted the teaching staff of the Water Management faculty in the development of these courses. We selected multiple experiments and found suitable locations for their execution, about a three-hour drive from the university. Several times we visited the fieldwork site together with students and teachers to explore the catchment, perform experiments, and gather data. For every experiment, a comprehensive manual was written with accompanying assignments and sheets tailor-made for the fieldwork excursion. All manuals are combined into a practical document that can be brought into the field.

The experiments require a wide variety of specific pieces of equipment. Most but not all of these were present at the university. With financial support from the Orange Knowledge Project (OKP), we were able to repair existing equipment and acquire new equipment that was lacking in order to facilitate the experiments that were deemed essential. We also re-organized part of the water lab where all equipment is stored to improve the equipment’s maintenance and organization.
With the completion of this project, we believe to have made a contribution to the improvement of the quality of education at HUNRE. We hope that a lot of students can benefit from our efforts as they go on the fieldwork excursion in the coming years.