Climate Reanalysis (SST) Forecasting
Time-series | Regression | RNN | U-Net | CNN | Regression | Numerical Data | Machine Learning
- Developed RNN and U-Net models for predicting Sea Surface Temperature (SST), sea currents, and salinity.
- Implemented CNN architectures for enhanced feature extraction from the East Sea and Yellow Sea data.
- Integrated numerical models like XGBoost, and Random Forest to optimize the forecast accuracy.
Reanalysis: (Retrospective Analysis), is a scientific method for developing a comprehensive record of how weather and climate are changing over time. A synthesized estimate of the state of the atmosphere, ocean and land by objectively merging observations and a numerical model information. A reanalysis typically extends over several decades or longer, and covers the entire globe or a region from the Earth’s surface to well above the stratosphere.
Sea surface temperature (SST) is a critical factor in the functioning of marine and terrestrial ecosystems, as well as human activities that rely on the ocean. Sea surface temperatures have risen by 0.5–0.6 °C since the 1950s, and over the oceans this has led to 4 percent more water vapor in the atmosphere since the 1970s. Sea surface temperature is measured by satellite instruments that record the energy emanating from the ocean surface globally. The energy is emitted at different wavelengths.
