Machine learning assisted hybrid models can improve streamflow simulation in diverse catchments across the conterminous US
Bibliographic Data
| ID | 15548165 |
|---|---|
| Authors | Goutam Konapala (0000-0003-0225-9633, Oak Ridge National Laboratory), Shih‐Chieh Kao (0000-0002-3207-5328, Oak Ridge National Laboratory, corresponding author), Scott Painter (0000-0002-0901-6987, Oak Ridge National Laboratory), Scott L Painter, Dan Lu (0000-0002-3000-5094, Oak Ridge National Laboratory) |
| Year | 2020 |
| Volume | 15 |
| Issue | 10 |
| Pages | 104022-104022 |
| Publication date | 2020-07-24 |
| Open Access | Yes |
| Type | ARTICLE |
| Venue | Environmental Research Letters (JOURNAL) |
| Journal identifiers | ISSN: 1748-9326 • E-ISSN: 1748-9326 |
| Publisher | IOP Publishing (PUBLISHER • GB) |
| DOI | 10.1088/1748-9326/aba927 |
| OpenAlex | W3044346761 |
| Language | EN |
| Citations received | 2 |
| References cited | 74 |
Incomplete representations of physical processes often lead to structural errors in process-based (PB) hydrologic models. Machine learning (ML) algorithms can reduce streamflow modeling errors but do not enforce physical consistency. As a result, ML algorithms may be unreliable if used to provide future hydroclimate projections where climates and land use patterns are outside the range of training data. Here we test hybrid models built by integrating PB model outputs with a ML algorithm known as Long Short-Term Memory (LSTM) network on their ability to simulate streamflow in 531 catchments representing diverse conditions across the Conterminous United States. Model performance of hybrid models as measured by Nash-Sutcliffe efficiency (NSE) improved relative to standalone PB and LSTM models. More importantly, hybrid models provide highest improvement in catchments where PB models fail completely (i.e., NSE < 0). However, all models performed poorly in catchments with extended low flow periods, suggesting need for additional research
Climatology · Consistency (knowledge bases · Drainage basin · Hydrological modelling · Hydrology (agriculture · Machine learning · Process (computing · Range (aeronautics · Streamflow · Computer Science · Environmental Science · Flood Risk Assessment and Management · Hydrological Forecasting Using AI · Hydrology and Watershed Management Studies · Artificial Intelligence · Geology
Variable Importance Assessment in Regression
The Parable of Google Flu
River flow forecasting through conceptual models part I — A discussion of principles
Definitions, methods, and applications in interpretable machine learning
Bias in random forest variable importance measures
Random Forests
Deep learning
Evaluation and machine learning improvement of global hydrological model-based flood simulations
| Citations per year | 0.4 |
|---|---|
| Citation span | 2021 - 2024 (4) |
| Highly cited | No |