Analysis of the Impact of Meteorological Factors Based on the MLP Correction Model on Long-wave Time Service System

As a critical ground-based navigation and time system, the propagation of long-wave timing signal is influenced by meteorological factors, specifically through two mechanisms: changes in atmospheric refractive index and variations in the surface conductivity characteristics along the propagation path. These factors affect propagation delay, thereby reducing timing and positioning accuracy. This paper analyzes the mechanisms by which meteorological conditions impact long-wave signal propagation and proposes a path delay correction model that integrates real-time meteorological data. Through correlation analysis of the propagation delay trend term ΔT with three meteorological factors—temperature, humidity, and pressure—the Spearman correlation coefficient between temperature and ΔT is 0.84, the Spearman correlation coefficient between humidity and ΔT is -0.51, both showing strong correlations, while the Spearman correlation coefficient between pressure and ΔT is -0.38, indicating a weak correlation. Further cluster analysis of ΔT under different temperature trends indicates that the influence of temperature on ΔT is static at the hourly scale. The model adjusts the path delay correction value using meteorological station data. It uses 600-hour ΔT data as the training set to predict 120-hour ΔT data multiple times. The test set shows that the timing error (STD, standard) improved from 27.55 ns before model correction to an average timing error (STD) of 10.48 ns after correction by MLP (multi-layer perceptron) model. The MLP correction model further reduced the average timing error (STD) to 10.05 ns. The results indicate that after correction by the MLP model, the timing error was reduced by 64%. This result has a positive impact on the performance improvement of long-wave time system in complex meteorological environments and provides some assistance for real-time and high-precision service assurance.