A note on leveraging synergy in multiple meteorological data sets with deep learning for rainfall

Section 1: Publication

Publication Type

Authorship

Kratzert, F., Klotz, D., Hochreiter, S., and Nearing, G. S.

Title

A note on leveraging synergy in multiple meteorological data sets with deep learning for rainfall–runoff modeling

Year

2021

Publication Outlet

Hydrol. Earth Syst. Sci., 25, 2685–2703

DOI

https://doi.org/10.5194/hess-25-2685-2021

ISBN

ISSN

Citation

Abstract

A deep learning rainfall–runoff model can take multiple meteorological forcing products as input and learn to combine them in spatially and temporally dynamic ways. This is demonstrated with Long Short-Term Memory networks (LSTMs) trained over basins in the continental US, using the Catchment Attributes and Meteorological data set for Large Sample Studies (CAMELS). Using meteorological input from different data products (North American Land Data Assimilation System, NLDAS, Maurer, and Daymet) in a single LSTM significantly improved simulation accuracy relative to using only individual meteorological products. A sensitivity analysis showed that the LSTM combines precipitation products in different ways, depending on location, and also in different ways for the simulation of different parts of the hydrograph.

Plain Language Summary

Section 2: Additional Information

Program Affiliations

GWF: Global Water Futures

Project Affiliations

GWF-CORE: Core Modelling and Forecasting

Submitters

Name	Role	Email	Institution

Publication Stage

Theme

Presentation Format

Additional Information