English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Over the past few decades, the hydrology community has witnessed notableadvancements in streamflow prediction, particularly with the introduction ofcutting-edge machine-learning algorithms. Recurrent neural networks, especiallyLong Short-Term Memory (LSTM) networks, have become popular due to theircapacity to create precise forecasts and realistically mimic the systemdynamics. Attention-based models, such as Transformers, can learn from theentire data sequence concurrently, a feature that LSTM does not have. This worktests the hypothesis that combining recurrence with attention can improvestreamflow prediction. We set up the Temporal Fusion Transformer (TFT)architecture, a model that combines both of these aspects and has never beenapplied in hydrology before. We compare the performance of LSTM, Transformers,and TFT over 2,610 globally distributed catchments from the recently availableCaravan dataset. Our results demonstrate that TFT indeed exceeds theperformance benchmark set by the LSTM and Transformers for streamflowprediction. Additionally, being an explainable AI method, TFT helps in gaininginsights into the streamflow generation processes.

Temporal Fusion Transformers for Streamflow Prediction: Value of Combining Attention with Recurrence