We used a deep learning method called long short-term memory (LSTM) to fill gaps in data collected by hydrologic monitoring networks. LSTM accounted for correlations in space and time and nonlinear trends in data. Compared to a traditional regression-based time-series method, LSTM performed comparably when filling gaps in data with smooth patterns, while it better captured highly dynamic patterns in data. Capturing such dynamics is critical for understanding dynamic complex system behaviors.

Here we test a maximum evaporation theory that acknowledges the interdependence between radiation, surface temperature, and evaporation over saturated land. We show that the maximum evaporation approach recovers observed evaporation and surface temperature under non-water-limited conditions across a broad range of bio-climates. The implication is that the maximum evaporation concept can be used to predict potential evaporation that has long been a major difficulty for the hydrological community.

We select the 2010–2012 UK drought and investigate an alternative unfolding of the drought from changes to its attributes. We created storylines of drier preconditions, alternative seasonal contributions, a third dry winter, and climate change. Storylines of the 2010–2012 drought show alternative situations that could have resulted in worse conditions than observed. Event-based storylines exploring plausible situations are used that may lead to high impacts and help stress test existing systems.

The role and impact of using additional data (other than runoff) for the prediction of daily hydrographs in ungauged basins are not well understood. In this study, we assessed the model performance in terms of runoff, soil moisture, and snow cover predictions with the existing regionalization approaches. Results show that the best transfer methods are the similarity and the kriging approaches. The performance of the transfer methods differs between lowland and alpine catchments.

We determined how snow and flow in small creeks change with temperature and precipitation in the Canadian Prairie, a region where water resources are often under stress. We tried something new. Every watershed in the region was placed in one of seven groups based on their landscape traits. We selected one of these groups and used its traits to build a model of snow and streamflow. It worked well, and by the 2040s there may be 20 %–40 % less snow and 30 % less streamflow than the 1980s.

Drought affects many regions worldwide, and future climate projections imply that drought severity and frequency will increase. Hence, the impacts of drought on the environment and society will also increase considerably. Monitoring and early warning systems for drought rely on several indicators; however, assessments on how these indicators are linked to impacts are still lacking. Our results show that meteorological indices are best linked to impact occurrences.

In this study, we investigated potentially catastrophic transitions in hydrological processes by identifying the early-warning signals which manifest as a critical slowing down in complex dynamic systems. We then analyzed the precipitation network of cities in the contiguous United States and found that key network parameters, such as the nodal density and the clustering coefficient, exhibit similar dynamic behaviour, which can serve as novel early-warning signals for the hydrological system.

This study presents a better characterization of surface hydrology variability in the Congo River basin, the second largest river system in the world. We jointly use a large record of in situ and satellite-derived observations to monitor the spatial distribution and different timings of the Congo River basin's annual flood dynamic, including its peculiar bimodal pattern.

In near-stream aquifers, mixing between stream water and ambient groundwater can lead to dilution and the removal of substances that can be harmful to the water ecosystem at high concentrations. We used a numerical model to track the spatiotemporal evolution of different water sources and their mixing around a stream, which are rather difficult in the field. Results show that mixing mainly develops as narrow spots, varying In time and space, and is affected by magnitudes of discharge events.

Rainfall erosivity is one of the main factors in soil erosion. A satellite-based global map of rainfall erosivity was constructed using data with a 30 min time interval. It was shown that the satellite-based precipitation products are an interesting option for estimating rainfall erosivity, especially in regions with limited ground data. However, ground-based high-frequency precipitation measurements are (still) essential for accurate estimates of rainfall erosivity.