DAILY PAPER REVIEW

20210218_Predicting lake surface water phosphorus dynamics using process-guided machine learning

1. Title, Journal and Authors

Title: Predicting lake surface water phosphorus dynamics using process-guided machine learning

Journal: Ecological Modelling

Authors: Paul C. Hanson1, Aviah B. Stillman1, Xiaowei Jia2, Anuj Karpatne3, Hilary A. Dugan, Cayelan C. Carey4, Joseph Stachelek5, Nicole K. Ward4, Yu Zhang6,7, Jordan S. Read8,Vipin Kumar

 

2. Summary

To investigate phosphorus concentration pattern in lake, the three models “Process-Guided Machine Mearning (PGML), Recurrent Neural Network(RNN), Process-Guided Recurrent Neural Network(PGRNN)” implement respectively.

PGML can’t explain that long-term trend and RNN can’t connect ecological principles. PGRNN was created by securing the shortcomings of these two and it was the best performance.

 

3. Originality and Creativity

This study compare actuality phosphorus load concentration in Lake Mendota and predicted by “PGML, RNN, PGRNN”. The phosphorus in lake was divided into Epilimnetic, Hypolimnetic, Sediment and calculated in a way that fits each other. Among the threee, Epilimnetic P is most important P that determine algal blooms and water quality. So this study focus on prediction of Epilimnetic P.

 

4. Application to research

Lumpping ecological knowledge and machine learning created a more accurate prediction model.

 

5. Contact

Ma, Kyoung Rim / M.S. program

 

Environmental Systems Engineering Lab.

School of Earth Sciences and Environmental Engineering

Gwangju Institute of Science and Technology

1 Oryong-dong Buk-gu Gwangju, 500-712, Korea

 

Phone : +82 10-5058-6832

E-mail : dkvmfhelxp34@gm.gist.ac.kr

첨부 (0)