This manuscript presents a timely and relevant study that assesses the performance of two gap-filling methods, DINEOF and DINCAE, applied to satellite-derived chlorophyll-a and phytoplankton functional type products in the Atlantic Ocean. The study is well-motivated, and the methodological rigor is commendable. The work contributes meaningfully to the ocean color remote sensing community and addresses a practical challenge in producing spatially continuous biogeochemical products. Therefore, I recommend this paper for publication. However, there are several aspects of the manuscript that would benefit from revision to improve clarity, structure, and academic presentation.

1. The manuscript is generally well-written, but there are a few areas where sentence structure and clarity could be improved. The introduction section is overly long and contains redundant background descriptions, which compromise the logical flow. It is recommended that the authors streamline this section by first clearly identifying the problem posed by data gaps in phytoplankton functional type products, followed by a focused overview of existing gap-filling methods, and then a concise rationale for selecting and comparing DINEOF and DINCAE in this study. A clearer structure will help readers better grasp the novelty and objectives of the work.
2. The detailed explanation of the DINEOF and DINCAE methods is commendable, providing insight into their respective advantages and limitations. While the methods section is detailed and provides useful information for reproducibility, the level of granularity is at times excessive, resembling a technical report rather than a scientific manuscript. Several parts describe operational or troubleshooting procedures in a narrative style that could be significantly condensed. For instance, the explanation regarding the use of the holdout method for generating the test dataset includes unnecessary implementation details that do not directly contribute to the scientific understanding. It is suggested to summarize such content more concisely, focusing on methodological rationale rather than process narration. Similarly, generic background statements—such as those explaining what hyperparameters are and why they matter—can be omitted or substantially shortened, with emphasis placed instead on the specific choices made in this study and their justification.
3. The results are well-presented, particularly the comparisons between DINEOF and DINCAE gap-filling methods. However, some of the claims about the relative advantages of DINCAE over DINEOF, especially in complex regions, could be better substantiated with more detailed statistical analysis. For instance, the statement that DINCAE outperforms DINEOF in specific regions (e.g., coastal areas) could benefit from a deeper analysis of the underlying causes for these discrepancies. While the manuscript effectively outlines the strengths and limitations of both gap-filling methods, it could further benefit from a more detailed discussion on the future improvements or modifications required for both techniques. Potential solutions to optimize computational efficiency could be discussed.
4. The study provides an important contribution to gap-filling in marine remote sensing, but the broader environmental implications of the findings could be more thoroughly explored. Specifically, how might these improved gap-filling techniques contribute to more accurate global phytoplankton monitoring, climate change studies, or marine conservation efforts?

The study entitled "Assessment of gap-filling techniques applied to satellite phytoplankton composition products for the Atlantic Ocean" by Mehdipour et al. addresses the critical challenge of missing data in satellite-derived ocean color products, particularly phytoplankton functional types (PFTs), which are crucial for understanding marine biogeochemical cycles and climate change impacts. The authors evaluate two established gap-filling methods, DINEOF and DINCAE, for reconstructing these datasets in the Atlantic Ocean. The manuscript is well written, interesting to read, the methodology sounds valid, and results and conclusions seem to be reasonable. I only have minor comments and suggestions for further improving the manuscript.

1. Page 8: the last two paragraphs are exactly the same (repeated).
2. DINEOF and DINCAE are different in terms of their input data. For DINCAE, only three days of data (previous and next day) are used to reconstruct the missing pixels. However, the DINEOF processing needs data of much longer time series as input. From Section 2.3.1, it is not clear how many days of data are used for input. I suggest that this important information included in the description. Also, please discuss the effect of the length of the time series on the performance of the two models.
3. Equation (1) is confusing: t is not defined, SST is not included, T is not specified, X-CHL is repeated for T times?
4. Section 3.2 and Fig. 5-6: Please discuss the effect of the data miss rate on the model performance in terms of different parameters (TChl vs. PFTs), and spatial regions (high latitude, equatorial Vs. mid-latitude).
5. Page 23, line 557: change VIIRS-SNAP to VIIRS-SNPP
6. It is noted that the PFTs shows significant difference with in situ measurement, but they are still valuable datasets. Further improvement of the quality of the PFT data are expected in the future.