Temporal Coherence Estimators for GBSAR

Round 1

Reviewer 1 Report

This paper presents a new estimator to express the temporal coherence for GBSAR. The authors build a theoretical relation between amplitude dispersion index and coherence, so promoting preliminary analysis, such as phase quality evaluation and coherent point selection, etc. This work is interesting, only several comments should be addressed before publication.

1.      The coherence estimation involves the ensemble average of samples in spatial domain, while amplitude dispersion index is calculated pixel-by-pixel. In this paper, the authors mentioned that the amplitude statistics was evaluated by averaging a group of pixels. But equation (16) does not contains a parameter of pixel number in spatial average. Does the spatial average number affect the relation between amplitude dispersion index and coherence? Please clarify.

2.      Both temporal coherence and amplitude dispersion index involve time-series estimation. Given the unchanged ground feature, the amplitude dispersion index approximates a stale value with the increase of temporal number, while the temporal coherence may suffer temporal decorrelation effect, which means the temporal coherence may follow a temporal model, e.g., exponential model or seasonal model. So how to evaluate the effect of temporal interval and acquisition number on the proposed relation in this study? Please clarify.

Author Response

see attached file

Author Response File: Author Response.pdf

Reviewer 2 Report

 Dear authors,
I read this paper with interest, and really appreciate his clarity and soundness. Nonetheless, the analysis of the results should be consolidated, with for instance a more detailed and specific presentation of each frequency (to avoid the overlap between the 3), and also with a quantitative assessment of the adequacy between the theoretical model and the experimental results. The latter should perhaps confirm the bias or skew effects which seem significant on figures 7 & 9. This additional analysis could also highlight the limitations of the Rician model, which should be better discussed and compare with other models more advanced that the Rayleigh one, in order to justify the very last statement about an easy generalization.

Author Response

see attached file

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have addressed my concerns.  I recommend publishing this paper: