**5. Conclusions**

This paper presents a classification methodology to distinguish coherent and incoherent DDMs in the CYGNSS land observations. Since the GNSS scattering signals from the windy ocean surface are almost incoherent, while the coherent land DDMs are closer to WAF, six different classification estimators are established based on scattering powerspreading shape and magnitude features over the ocean, and land CYGNSS collected DDMs, which are used to screen the land high confidence coherent component dominated DDMs. The results show that the estimators based on the absolute magnitude features of DDM are difficult to distinguish its coherency, while the estimator indicating shape features performs better. The average proportion of GNSS-R land observations dominated by coherent components is 89.6%. NIDW-derived TES performs best among all defined DDM observables, and its PDFs from the ocean and land DDMs are more separated and sharper, whose detection probability for coherent observations can reach 93.8% with the lowest detection probability of error. The distribution of high-confidence coherent and incoherent surface observation indicates that observations over the dense forest cannot change the surface scattering properties but will greatly weaken the coherent scattering power. Using 19 months of CYGNSS observation data and SMAP SM product for land SM retrieval model validation, the RMSE of model performance with k-fold cross-validation can reach 0.04 cm3/cm3. Incoherent observations have not seriously impaired the accuracy of CYGNSS soil moisture inversion.

**Author Contributions:** S.J. and Z.D. conceived and designed the experiments and Z.D. performed the experiments and analyzed the data. Both authors contributed to the writing of the paper. Both authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Strategic Priority Research Program Project of the Chinese Academy of Sciences (Grant No. XDA23040100), and Shanghai Leading Talent Project (Grant No. E056061).

**Acknowledgments:** Authors thank NASA for providing CYGNSS and SMAP data.

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