Physics-guided architecture (PGA) of LSTM models for uncertainty quantification in lake temperature modeling
Links
- More information: Publisher Index Page (via DOI)
- Download citation as: RIS | Dublin Core
Abstract
This chapter focuses on meeting the need to produce neural network outputs that are physically consistent and also express uncertainties, a rare combination to date. It explains the effectiveness of physics-guided architecture - long-short-term-memory (PGA-LSTM) in achieving better generalizability and physical consistency over data collected from Lake Mendota in Wisconsin and Falling Creek Reservoir in Virginia, even with limited training data. Even though PGL formulations result in improvements in the generalization performance and lead to machine learning (ML) predictions that are more physically consistent, simply adding the physics-based loss function in the learning objective does not overcome the black-box nature of neural network architectures, which often involve arbitrary design choices. The temperature of water in a lake is a fundamental driver of lake biogeochemical processes, and it controls the growth, survival, and reproduction of fishes in the lake.
Study Area
| Publication type | Book chapter |
|---|---|
| Publication Subtype | Book Chapter |
| Title | Physics-guided architecture (PGA) of LSTM models for uncertainty quantification in lake temperature modeling |
| Chapter | 17 |
| DOI | 10.1201/9781003143376-17 |
| Year Published | 2022 |
| Language | English |
| Publisher | Taylor & Francis |
| Contributing office(s) | WMA - Integrated Information Dissemination Division |
| Description | 18 p. |
| Larger Work Type | Book |
| Larger Work Subtype | Monograph |
| Larger Work Title | Knowledge-guided machine learning: Accelerating discovery using scientific knowledge and data |
| First page | 399 |
| Last page | 416 |
| Country | United States |
| State | Virginia, Wisconsin |
| Other Geospatial | Falling Creek Reservoir, Lake Mendota |
| Google Analytic Metrics | Metrics page |