.png){kind=icon}
.png){kind=icon}
Return
Uncertainty-informed topology-aware transfer learning for robust stream salinity prediction
.png){kind=icon}
OA
AI
DOI：10.1016/j.ejrh.2026.103487.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Used topology-aware transfer learning for stream salinity prediction. • Generated robust synthetic stream salinity data for the Upper Red River Basin, US. • Quantified modeling uncertainty through prediction intervals via the LUBE method. • Evaluated the salinity prediction framework via continuous sub-daily datasets.
Keywords:
CWC
Coverage Width-based Criterion
FFNN
Feedforward Neural Network
LUBE
Lower Upper Bound Estimation
ML
Machine learning
MPIW
Mean Prediction Interval Width
NSE
Nash Sutcliffe Efficiency
NMPIW
Normalized Mean Prediction Interval Width
OWRB
Oklahoma Water Resources Board
OWRC
Oklahoma Water Resources Center
PI
Prediction Interval
PICP
Prediction Interval Coverage Probability
RMSE
Root Mean Square Error
SC
Specific Conductance
TL
Transfer learning
TL-NF
Transfer learning without freezing layers
TL-FR
Transfer learning with freezing layers
USGS
United States Geological Survey
URRB
Upper Red River Basin
Water salinity
Machine learning
Missing data
Intrabasin transfer learning
Uncertainty quantification
Upper red river basin
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
J
IF:
0
Papers:
604
Citations:
0