| Title |
Applications of Machine Learning Models for the Estimation of Reservoir CO2 Emissions |
| Authors |
유지수 ( Jisu Yoo ) ; 정세웅 ( Se-woong Chung ) ; 박형석 ( Hyung-seok Park ) |
| DOI |
https://doi.org/10.15681/KSWE.2017.33.3.326 |
| Keywords |
Artificial neural network; Carbon emission; Daecheong Reservoir，Machine learning; Random forest |
| Abstract |
Lakes and reservoirs have been reported as significant sources of carbon emissions released into the atmosphere of many countries. Although field experiments and theoretical investigations based on the fundamental gas exchange theory have proposed quantitative amounts of Net Atmospheric Flux (NAF) in various climate regions, there is significant uncertainty about the global scale estimation. Mechanistic models can be manipulated for understanding and estimating temporal and spatial variations of NAFs for considering complex hydrodynamic and biogeochemical processes in a reservoir, but such models require extensive and costly datasets and model parameters. However, data driven machine learning (ML) algorithms are likely to be alternative tools to estimate NAFs in responding to independent environmental variables. The objective of this study was to develop random forest (RF) and multi-layer artificial neural network (ANN) models for the estimation of the daily CO2 NAFs in Daecheong Reservoir located in Geum River in South Korea, and compare the models` performance against the multiple linear regression (MLR) model that was proposed in the previous study (Chung et al” 2016). As a result, the RF and ANN models revealed much enhanced performance in the estimation of high NAF values, while the MLR model significantly underestimated them. A cross-validation with 10-fold random samplings was applied to evaluate the performance of the three models，and it indicated that the ANN model is best, followed by RF and MLR models. |