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

Accurate and efficient models for rainfall runoff (RR) simulations arecrucial for flood risk management. Most rainfall models in use today areprocess-driven; i.e. they solve either simplified empirical formulas or somevariation of the St. Venant (shallow water) equations. With the development ofmachine-learning techniques, we may now be able to emulate rainfall modelsusing, for example, neural networks. In this study, a data-driven RR modelusing a sequence-to-sequence Long-short-Term-Memory (LSTM) network wasconstructed. The model was tested for a watershed in Houston, TX, known forsevere flood events. The LSTM network's capability in learning long-termdependencies between the input and output of the network allowed modeling RRwith high resolution in time (15 minutes). Using 10-years precipitation from153 rainfall gages and river channel discharge data (more than 5.3 million datapoints), and by designing several numerical tests the developed modelperformance in predicting river discharge was tested. The model results werealso compared with the output of a process-driven model Gridded SurfaceSubsurface Hydrologic Analysis (GSSHA). Moreover, physical consistency of theLSTM model was explored. The model results showed that the LSTM model was ableto efficiently predict discharge and achieve good model performance. Whencompared to GSSHA, the data-driven model was more efficient and robust in termsof prediction and calibration. Interestingly, the performance of the LSTM modelimproved (test Nash-Sutcliffe model efficiency from 0.666 to 0.942) when aselected subset of rainfall gages based on the model performance, were used asinput instead of all rainfall gages.

High Temporal Resolution Rainfall Runoff Modelling Using Long-Short-Term-Memory (LSTM) Networks