A PHP Error was encountered

Severity: Warning

Message: file_get_contents(https://...@gmail.com&api_key=61f08fa0b96a73de8c900d749fcb997acc09&a=1): Failed to open stream: HTTP request failed! HTTP/1.1 429 Too Many Requests

Filename: helpers/my_audit_helper.php

Line Number: 197

Backtrace:

File: /var/www/html/application/helpers/my_audit_helper.php
Line: 197
Function: file_get_contents

File: /var/www/html/application/helpers/my_audit_helper.php
Line: 271
Function: simplexml_load_file_from_url

File: /var/www/html/application/helpers/my_audit_helper.php
Line: 3165
Function: getPubMedXML

File: /var/www/html/application/controllers/Detail.php
Line: 597
Function: pubMedSearch_Global

File: /var/www/html/application/controllers/Detail.php
Line: 511
Function: pubMedGetRelatedKeyword

File: /var/www/html/index.php
Line: 317
Function: require_once

Innovative deep learning and signal decomposition approaches for enhanced spatial and temporal suspended sediment concentration prediction. | LitMetric

Category Ranking

98%

Total Visits

921

Avg Visit Duration

2 minutes

Citations

20

Article Abstract

Accurate prediction of suspended sediment concentration (SSC) is critical for effective river basin planning, optimal water resource management, and successful ecological restoration. This study introduces novel hybrid deep learning approaches, utilizing long short-term memory (LSTM), bidirectional long short-term memory (BiLSTM), gated recurrent unit (GRU), and dense deep neural networks (DDNNs) in combination with complementary ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and successive variational mode decomposition (SVMD). Initially, to determine the most influential parameters, a wavelet transform coherence (WTC) analysis was performed, which indicated that turbidity, gauge height, water temperature, discharge, and specific conductance had the strongest correlations with SSC. The model with all identified input parameters was trained using datasets collected from four sequential gauges along the Lower Colorado River in the USA, covering the period from 2008 to 2022. A comparative analysis of standalone and integrated deep learning models demonstrated that standalone DDNN and CEEMDAN-DDNN and SVMD-DDNN hybrid models outperformed others across all gauges. Results revealed that SVMD was more effective than CEEMDAN in enhancing model efficiency, achieving the highest prediction accuracy at the third gauge, with a mean square error of 0.027, root mean square error of 0.165, Nash-Sutcliffe efficiency of 0.983, and R of 0.992. Furthermore, the SSC values from 6 days prior at upstream gauges had a significant impact on SSC prediction in the downstream gauge. Utilizing the SVMD-DDNN hybrid model increased prediction accuracy by 3.8%. Moreover, density plot analysis revealed that the CEEMDAN-DDNN and SVMD-DDNN hybrid models demonstrated the best distribution matching with observed SSC values at gauge 3. These advanced frameworks offer a precise and reliable solution for SSC prediction, significantly enhancing river management practices.

Download full-text PDF

Source
http://dx.doi.org/10.1007/s11356-025-36581-3DOI Listing

Publication Analysis

Top Keywords

deep learning
12
svmd-ddnn hybrid
12
suspended sediment
8
sediment concentration
8
long short-term
8
short-term memory
8
mode decomposition
8
models demonstrated
8
ceemdan-ddnn svmd-ddnn
8
hybrid models
8

Similar Publications