Underwater three-dimensional temperature-salinity parallel forecasting method based on LightGBM model

Abstract

The invention discloses an underwater three-dimensional temperature-salinity parallel forecasting method based on a LightGBM model, and relates to the technical field of ocean. The method comprises the following steps: carrying out data preprocessing on obtained sea surface temperature historical data, sea surface height historical data and Argo buoy historical data; and after a sea surface temperature abnormal value, a sea surface height abnormal value and an Argo buoy abnormal value which are uniform in temporal-spatial resolution are obtained through screening, constructing a training data set, and performing parallel training based on the LightGBM model to obtain an underwater three-dimensional temperature-salinity forecasting model for underwater three-dimensional temperature-salinity forecasting. The method fully utilizes rich sea surface satellite data, and realizes analysis and forecasting of the underwater three-dimensional temperature-salinity information through a parallel algorithm, so that the forecasting precision of the three-dimensional temperature-salinity information is guaranteed, the forecasting efficiency is remarkably improved, and the limitation of inversion of the underwater three-dimensional temperature-salinity information in dynamics and statistics is broken through.

Description

technical field [0001] The invention relates to the field of marine technology, in particular to an underwater three-dimensional temperature and salinity parallel forecasting method based on the LightGBM model. Background technique [0002] Underwater platform mission planning, concealed or safe navigation, sonar hydroacoustic detection, ship navigation, oil platforms, and maritime rescue all rely on large-scale, accurate three-dimensional ocean temperature and salinity information. Due to the huge manpower and financial resources required for offshore surveys and experiments, and the number of Argo buoys is still relatively small for the entire ocean, the underwater three-dimensional temperature and salinity information is lacking or even almost non-existent. At present, the methods for analyzing and calculating the three-dimensional temperature and salinity information of the ocean mainly include dynamical methods or statistical analysis methods. However, dynamical methods...

AI Technical Summary

Problems solved by technology

Due to the huge manpower and financial resources required for sea surveys and experiments, and the number of Argo buoys is still relatively small for the entire ocean, the underwater three-dimensional temperature and salinity information is lacking or even almost non-existent.

At present, the methods for analyzing and calculating the three-dimensional temperature and salinity information of the ocean mainly include dynamical methods or statistical analysis methods. However, dynamical methods are black-box and difficult to implement and require a large amount of calculation. Statistical analysis methods rely on long-term historical data and ocean development. Chaos is difficult to capture, so the research on three-dimensional temperature and salinity forecasting is insufficient

Images