Long‐Term Prediction Model for Erosion‐Deposition Topographic Evolution in the Sanmenxia‐To‐Xiaolangdi Reach of the Yellow River Based on Deep Learning
Bibliographic record
Abstract
Abstract Reservoirs are essential for global water management and energy regulation, but sedimentation threatens their longevity. This study investigates a 130 km section of the Yellow River between the Sanmenxia and Xiaolangdi dams, using deep learning to predict long‐term erosion and deposition patterns. From 2009 to 2023, we gathered water depth data from 56 sites (840 measurements) with unmanned survey boats and drone‐based LiDAR (Light Detection and Ranging), alongside flow and sediment records. After preprocessing, we evaluated three machine learning models: Convolutional Network for Multimodal Time Series (CNN‐MTS), Convolutional Transformer for Multimodal Time Series (CNN‐Transformer‐MTS), and Convolutional Bi‐LSTM for Multimodal Time Series (CNN‐BiLSTM‐MTS). The CNN‐BiLSTM‐MTS model excelled, achieving a mean absolute error (MAE) of 17.84 m, a coefficient of determination ( R 2 ) of 0.9916, and reducing errors by up to 26% compared to alternatives. Key drivers of sediment dynamics included sediment load, maximum sediment concentration, and maximum flow. Data from 2009 to 2023 showed elevation shifts from −0.21 m near the dam to +1.158 m at the reservoir's tail. Predictions for 2024 to 2050 suggest varied riverbed changes, with the Guxian Reservoir's operation in 2036 expanding elevation ranges from −0.625 to 0.875 m. These findings highlight deep learning's potential for efficient sediment management in reservoirs and offer insights for sustainable hydraulic engineering. However, uncertainties persist in scaling the model, improving data resolution, and coordinating across regions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".