**5. Conclusions**

In this study, RFI-affected AMSR-2 C-band data regarding the U.S. land area in 2016 were accurately repaired through iterative principal component analysis (PCA). The STD and bias characteristics of the brightness temperature data in the C-band vertical polarization channel were compared and analyzed before and after restoration to verify the assimilation potential of the repaired data. The main conclusions of this work are described below.

(1) The NPCA method was used to identify RFI signals in the observed brightness temperature data representing the U.S., collected from the 6.9-GHz channel for 2016. The results showed that severe RFI impacts persisted throughout the year in the U.S. The interference sources were mainly distributed in areas containing cities, such as the states of Virginia, North Carolina, and Texas. The amount of data suffering from RFI accounted for approximately 40% of the total amount of analyzed data.

(2) Based on the iterative PCA method applied herein, the disturbed brightness temperatures throughout the year were repaired. On the whole, the abnormally high brightness temperatures corresponding to RFI areas were repaired with a high level of precision. The overall brightness temperature distribution conformed to natural surface emission characteristics, maintaining good spatial continuity following the repair process, with small-scale features also being effectively recovered. At the same time, the applied restoration method was not affected by seasonal changes in brightness temperature or by variations in terrain or vegetation types and thus exhibited good stability and prospects for long-term RFI data recovery.

(3) The STD and bias in RFI-affected areas were significantly reduced following the restoration process; in addition, both of them were consistent with the corresponding values obtained from the pollution-free data, indicating that the repaired data retained the bias and STD characteristics of the observation instrument. Furthermore, in pine-forestand brush-covered areas, the restoration method had an obvious improvement effect. Over land, the STD decreased gradually with increasing terrain, but the trend of the bias was the opposite. These findings will be useful for subsequent data assimilation applications.

**Author Contributions:** Conceptualization, Z.Q. and W.S.; methodology, W.S.; software, W.S.; validation, Z.Q., W.S. and Z.L.; formal analysis, Z.L.; investigation, X.B.; resources, Z.Q.; data curation, W.S.; writing—original draft preparation, W.S.; writing—review and editing, Z.Q.; visualization, W.S.; supervision, Z.L.; project administration, Z.Q.; funding acquisition, Z.Q. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was jointly funded by the National Key R&D Program of China (Grant 2018YFC1507302), the National Natural Science Foundation of China (Grant 42075166), the Youth Project of the National Natural Science Foundation of China (41805076), the Natural Science Foundation of Jiangsu Province (BK20211396) and the FengYun-3 meteorological satellite engineering ground application Project (FY-3(03)-AS-11.08).

**Data Availability Statement:** ERA5 hourly reanalysis data: https://cds.climate.copernicus.eu/ cdsapp#!/search?type=dataset&text=era5 (accessed on 2 March 2022); AMSR-2 brightness temperature data: https://gportal.jaxa.jp/gpr/search?tab=0 (accessed on 13 July 2021).

**Acknowledgments:** We would like to acknowledge the suggestions given by reviewers and editor. And we are grateful to the High-Performance Computing Centre of the Nanjing University of Information Science and Technology for the performed numerical calculations in this study using its blade cluster system.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

