**5. Conclusions**

The surface energy balance of the polar region is mainly dependent upon the surface albedo characteristics. Ice-albedo feedback can cause substantial alteration of absorbed solar energy with even slight fluctuations in albedo; therefore, the demand for fine-resolution satellite sea-ice albedo is increasing. We worked on developing an operational VIIRS sea-ice albedo (VSIA) product and validated the product with nearly six years of in situ measurements at 30 automatic weather stations from PROMICE and GC-NET.

The direct estimation algorithm used in the VSIA product has unique advantages for satellite albedo retrieval. First, it does not need to collect the multi-angle dataset as an input for real-time processing. Second, it generated blue-sky albedo values that can be directly compared with ground measurements.

We used two sources of ground measurements as reference data for VSIA validation. The first is the long-term AWS observed albedo over Greenland. These datasets have high quality and large samples. The comparison reveals good agreemen<sup>t</sup> between the VSIA and PROMICE observations with a bias of 0.028 and RMSE of 0.072—a comparable result with historical validation results from the snow-specific algorithm in the Greenland area. At some large SZA conditions, the VSIA uncertainty shows a slight increasing trend, which is a result of combined uncertainties in the LUT precision, observation uncertainty, model accuracy, and spatial heterogeneity influence. The limitation of using the AWS albedo measurements is that the dataset cannot represent the surface conditions other than the optically thick sea ice. Therefore, we also used the ground measurements from previous field experiments over the sea-ice surface as the reference data. The comparison of VSIA with these datasets suffered from many issues including the limited sample size, scale difference, and strong spatial heterogeneity. Still, the VSIA retrievals match the ground measurements in magnitude and roughly reflect the evolution trend. The bias between the retrieved instantaneous albedo and measured albedo is 0.09 in the central Arctic dataset, and the bias between the retrieved daily mean albedo with the measured value is 0.077 in the Alaska dataset. Future continuous evaluation attempts are planned for fully understanding VSIA accuracy, including cross-comparison with other available sea-ice albedo products and direct-validation using more ground measurements.

With the completion of the S-NPP VIIRS granule albedo development, we now have the integration plan of a 1 km gridded global surface albedo product. The gridded albedo product will be map-projected and convenient to overlay with other data sets in applications. Considering the daily composite requirement for a gridded albedo product, we will develop a LUT directly linking TOA reflectance to the daily mean albedo of sea ice. Daily mean albedo will consider the diurnal variation pattern of sea-ice albedo so that the values from different collection times can be calculated together. Moreover, the NOAA-20 (designated as JPSS-1) was launched on 18 November 2017, with the VIIRS carried aboard, and joined the S-NPP satellite in the same orbit. The granule/gridded VSIA albedo will also be produced using NOAA-20 observations.

**Author Contributions:** Conceptualization, Y.Y and J.P.; Methodology, J.P.; Validation, J.P.; Formal Analysis, J.P.; Writing-Original Draft Preparation, J.P.; Writing-Review & Editing, Y.Y., P.Y., and S.L.; Funding Acquisition, Y.Y.".

**Funding:** This study was supported by NOAA gran<sup>t</sup> NA14NES4320003 (Cooperative Institute for Climate and Satellites -CICS) at the University of Maryland/ESSIC.

**Acknowledgments:** This study was supported by JPSS program at NOAA/NESDIS Center for Satellite Applications and Research, for the VIIRS albedo product development. The manuscript contents are solely the opinions of the authors and do not constitute a statement of policy, decision, or position on behalf of NOAA or the U.S. Government. The authors would like to thank Dr. Xiwu Zhan and Dr. Jeffrey R. Key for helpful comments on earlier drafts of the manuscript. Many thanks to Mr. Lizhao Wang for precious suggestions and data

support for the study. They are also grateful to Mr. Joshua Hrisko for language editing. We thank two anonymous reviewers for their careful reading of our manuscript and their many insightful comments and suggestions on earlier drafts of the manuscript.

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