In this paper, we present the results of the “Rate My Hydrograph” study, where we compare expert ratings of simulated hydrographs with quantitative metrics.
Posts
In this paper, we present the full operational framework used by the Flood Forecasting team at Google.
This paper performs a rigorous benchmark of traditional hydrologic models and an LSTM-based model for rainfall-runoff modeling.
This paper introduces the Caravan dataset, a global large-sample hydrology dataset that builds on cloud computing to be extensible by anyone.
Oral presentation at the EGU General Assembly 2022 on a social study to compare expert rankings of simulated hydrographs with quantitative metrics.
Accompanying paper to our open source Python library NeuralHydrology.
This paper investigates the hypothesis that the lack of enforced mass conservation is the main reason that deep learning models outperform traditional hydrology models.
Presentation at the AGU 2021 Fall Meeting presenting approaches to improve LSTM streamflow predictions with near-real-time observation data.
In this paper, we investigate the potential of using the LSTM as a post-processor for the US National Water Model.
Technical note that compares autoregression to data assimilation for deep learning models and rainfall-runoff modeling.
This paper investigates the hypothesis that deep learning models may not be reliable in extrapolating extreme events.
In this paper we show the benefits of using multiple meteorological forcing products at the same time in a single LSTM-based rainfall-runoff model over just using a single product.
Oral presentation at the virtual EGU General Assembly 2021 on rainfall–runoff forecasting with Multi-Timescale LSTM.
Oral presentation at the virtual EGU General Assembly 2021 on rainfall–runoff prediction with Graph Neural Networks.
Oral presentation at the virtual EGU General Assembly 2021 on uncertainty prediction with LSTMs in the context of rainfall-runoff modeling.
Deep learning based uncertainty estimation techniques and benchmarking procedure for rainfall-runoff modeling.
LamaH-CE contains a collection of runoff and meteorological time series as well as various (catchment) attributes for 859 gauged basins in the upper Danube catchment and Austria.
In this study, we present a mass-conserving variant of the LSTM and its application to arithmetic tasks, traffic forecasting, modeling a pendulum and rainfall-runoff modeling.
Oral presentation at the virtual AGU Fall Meeting 2020 on streamflow prediction at arbitraty timescales with a single LSTM-based model.
Oral presentation at the virtual AGU Fall Meeting 2020 on uncertainty prediction with LSTMs in the context of rainfall-runoff modeling.
Spotlight talk at the AI for Earth Sciences workshop of the NeurIPS 2020, presenting introduction of the world and terminology of hydrology/streamflow prediction for data scientists.
Spotlight talk at the AI for Earth Sciences workshop of the NeurIPS 2020, presenting a first glimpse of a new LSTM-based model that conserves mass by design: the Mass-Conserving LSTM
Opinion paper discussing the future of Hydrology, especially in the context of recent developments in Machine Learning
Invited talk at the EGU General Assembly 2020 on using LSTMs for flood forecasting in ungauged basins.
Virtual presentation at the EGU General Assembly 2020 comparing LSTMs trained for each basin individually with a single LSTM trained for all basins together.
Virtual presentation at the EGU General Assembly 2020 presenting the HydroNets architecture for modeling entire river networks.
Modeling entire nested river trees by integrating the river hierachy into the neural network architecture. This manuscripts proposes HydroNets, an architecture designed for modeling multiple nested gauge stations.
Workshop paper, investigating first ways of using LSTM-based models for climate change related questions in hydrology.
Workshop paper, investigating first ways of using LSTM-based models for climate change related questions in hydrology.
Poster at the AGU 2019 Fall Meeting presenting Martins work on a model comparison study for the Great Lakes/Lake Erie area.
Presentation at the AGU 2019 Fall Meeting presenting our recent results on large-scale hydrological modeling using LSTMs.
In this manuscript we test LSTM-based rainfall-runoff models on the task of prediction in ungauged basins and show, that a single LSTM-based model does better prediction in ungauged basins than a traditional hydrological model that was specifically calibrated for each basin individually.
This paper investigates the influence of the number of training basins and the training period length on the model performance for the EA-LSTM and XGBoost
In this manuscript we show for the first time how to train a single LSTM-based neural network as general hydrology model for hundreds of basins. Furthermore, we proposed the Entity-Aware LSTM (EA-LSTM) in which static features are used explicitly to subset the model for a specific entity (here a catchment).