Predicting surface temperature in Lake Villarrica (Chilean Patagonia) using a long short-term memory model

Area of Science:

• Environmental Science
• Data Science
• Remote Sensing

Background:

• Water temperature trends in Lake Villarrica show warming, particularly in spring and summer.
• Long-term surface temperature data (1986-2020) reveal annual maximums and minimums.
• Inland water quality monitoring is crucial for environmental management.

Purpose of the Study:

• To analyze long-term water temperature trends in Lake Villarrica.
• To develop and evaluate a machine learning model for predicting lake surface temperatures.
• To integrate remote sensing and in situ data for enhanced water monitoring.

Main Methods:

• Analysis of 34-year water temperature time series data.
• Development of a machine learning model using physical, chemical, biological, meteorological, and satellite spectral data.
• Application of the Long Short-Term Memory (LSTM) model for temperature estimation.

Main Results:

• Spring and summer surface temperatures in Lake Villarrica have increased over the study period.
• The LSTM model demonstrated high accuracy in estimating surface temperatures across the lake.
• The west region of the lake showed the best model performance with low error metrics (RMSE=2.79, MedAE=2.13).

Conclusions:

• Machine learning, particularly LSTM, is effective for monitoring lake surface temperatures.
• The integration of remote sensing and machine learning offers a powerful approach for inland water management.
• Continued monitoring and advanced modeling are essential for understanding and managing aquatic ecosystems.