Using Long Short-Term Memory networks to connect water table depth anomalies to precipitation anomalies over Europe

Ma, Yueling; Montzka, Carsten; Bayat, Bagher; Kollet, Stefan

doi:10.5194/hess-2020-382

Cited by 2 publications

(1 citation statement)

References 41 publications

(60 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This "big" dataset is also potentially useful for developing/experimenting with the data-driven method because of its high volume (e.g., long time series and high resolution) and variety (e.g., fully interactive states and fluxes). For example, Ma et al (2020Ma et al ( , 2021 used TSMP-G2A to extract long-term memory correlations using deep learning to predict groundwater table depth anomalies, using precipitation and soil moisture information.…”

Section: Study Area and Datamentioning

confidence: 99%

Deep learning of model- and reanalysis-based precipitation and pressure mismatches over Europe

Yousefi

Kollet

2023

Front. Water

Self Cite

View full text Add to dashboard Cite

Physically based numerical weather prediction and climate models provide useful information for a large number of end users, such as flood forecasters, water resource managers, and farmers. However, due to model uncertainties arising from, e.g., initial value and model errors, the simulation results do not match the in situ or remotely sensed observations to arbitrary accuracy. Merging model-based data with observations yield promising results benefiting simultaneously from the information content of the model results and observations. Machine learning (ML) and/or deep learning (DL) methods have been shown to be useful tools in closing the gap between models and observations due to the capacity in the representation of the non-linear space–time correlation structure. This study focused on using UNet encoder–decoder convolutional neural networks (CNNs) for extracting spatiotemporal features from model simulations for predicting the actual mismatches (errors) between the simulation results and a reference data set. Here, the climate simulations over Europe from the Terrestrial Systems Modeling Platform (TSMP) were used as input to the CNN. The COSMO-REA6 reanalysis data were used as a reference. The proposed merging framework was applied to mismatches in precipitation and surface pressure representing more and less chaotic variables, respectively. The merged data show a strong average improvement in mean error (~ 47%), correlation coefficient (~ 37%), and root mean square error (~22%). To highlight the performance of the DL-based method, the results were compared with the results obtained by a baseline method, quantile mapping. The proposed DL-based merging methodology can be used either during the simulation to correct model forecast output online or in a post-processing step, for downstream impact applications, such as flood forecasting, water resources management, and agriculture.

show abstract

Section: Study Area and Datamentioning

confidence: 99%

Deep learning of model- and reanalysis-based precipitation and pressure mismatches over Europe

Yousefi

Kollet

2023

Front. Water

Self Cite

View full text Add to dashboard Cite

show abstract

Using Long Short-Term Memory networks to connect water table depth anomalies to precipitation anomalies over Europe

Ma¹,

Montzka²,

Bayat³

et al. 2021

Hydrol. Earth Syst. Sci.

Self Cite

View full text Add to dashboard Cite

Abstract. Many European countries rely on groundwater for public and industrial water supply. Due to a scarcity of near-real-time water table depth (wtd) observations, establishing a spatially consistent groundwater monitoring system at the continental scale is a challenge. Hence, it is necessary to develop alternative methods for estimating wtd anomalies (wtda) using other hydrometeorological observations routinely available near real time. In this work, we explore the potential of Long Short-Term Memory (LSTM) networks for producing monthly wtda using monthly precipitation anomalies (pra) as input. LSTM networks are a special category of artificial neural networks that are useful for detecting a long-term dependency within sequences, in our case time series, which is expected in the relationship between pra and wtda. In the proposed methodology, spatiotemporally continuous data were obtained from daily terrestrial simulations of the Terrestrial Systems Modeling Platform (TSMP) over Europe (hereafter termed the TSMP-G2A data set), with a spatial resolution of 0.11∘, ranging from the years 1996 to 2016. The data were separated into a training set (1996–2012), a validation set (2013–2014), and a test set (2015–2016) to establish local networks at selected pixels across Europe. The modeled wtda maps from LSTM networks agreed well with TSMP-G2A wtda maps on spatially distributed dry and wet events, with 2003 and 2015 constituting drought years over Europe. Moreover, we categorized the test performances of the networks based on intervals of yearly averaged wtd, evapotranspiration (ET), soil moisture (θ), snow water equivalent (Sw), soil type (St), and dominant plant functional type (PFT). Superior test performance was found at the pixels with wtd < 3 m, ET > 200 mm, θ>0.15 m3 m−3, and Sw<10 mm, revealing a significant impact of the local factors on the ability of the networks to process information. Furthermore, results of the cross-wavelet transform (XWT) showed a change in the temporal pattern between TSMP-G2A pra and wtda at some selected pixels, which can be a reason for undesired network behavior. Our results demonstrate that LSTM networks are useful for producing high-quality wtda based on other hydrometeorological data measured and predicted at large scales, such as pra. This contribution may facilitate the establishment of an effective groundwater monitoring system over Europe that is relevant to water management.

show abstract

Using Long Short-Term Memory networks to connect water table depth anomalies to precipitation anomalies over Europe

Cited by 2 publications

References 41 publications

Deep learning of model- and reanalysis-based precipitation and pressure mismatches over Europe

Deep learning of model- and reanalysis-based precipitation and pressure mismatches over Europe

Using Long Short-Term Memory networks to connect water table depth anomalies to precipitation anomalies over Europe

Contact Info

Product

Resources

About