Objective The FY-4A surface solar irradiance (SSI) product provides high spatiotemporal-resolution solar radiation data. Validating and effectively correcting this product is critical for accurate solar energy resource assessment, supporting the achievement of the "carbon peaking" and "carbon neutrality" goals.

Method Based on the data from five ground radiation observation stations in Hubei Province from 2019 to 2020, and compared with the satellite-derived total irradiance during the same period, two methods, namely linear regression and probability density function (PDF) matching, were employed to correct the hourly total irradiance of FY-4A and compare their effects.

Result The results show that: FY-4A-derived irradiance was significantly correlated with the observed values, with a bias of about 190 W/m². The mean bias showed a temporal distribution opposite to that of the total irradiance itself—that is, the smaller the total irradiance, the larger the bias. In the annual cycle, the bias was largest in January and smallest in August; in the diurnal cycle, the bias was smallest at 08:00 and largest at 16:00. For both correction methods, the correlation coefficients showed no significant change before and after correction, but the seasonal mean absolute error decreased by 40%–60%, and the mean bias was reduced from 149.7–247.9 W/m² (before correction) to −30–20 W/m² (after correction).

Conclusion  FY-4A SSI exhibits strong correlation with ground observations in Hubei, and systematic biases can be substantially reduced through correction. The linear regression and probability density matching (PDF) methods demonstrated comparable effectiveness in correcting biases during spring and summer, while the PDF method outperformed in autumn and winter.