The review article ‘Synthesis of Data Products for Ocean Carbonate Chemistry’ offers a thorough and valuable summary of existing ocean carbonate data products, which will greatly benefit the scientific community. However, it overlooks two machine learning-based products that focus on improving surface pCO2 estimates in the Indian Ocean region. 

The first is an ML-based climatological pCO2 data product recently developed and published in the journal Scientific Data for the Bay of Bengal region (https://www.nature.com/articles/s41597-024-03236-w) (Joshi et al., 2024). This data product integrates publicly available open-ocean observations with data from the Indian Exclusive Economic Zone. Given that the Bay of Bengal is a unique basin with very limited publicly accessible pCO2 observations, this high-resolution (~0.083°) climatological pCO2 data product represents a significant advancement in our understanding of pCO2 dynamics in the region. Therefore, it may be appropriate to include this product in Section 3.1.3 (i.e., Gridded and derived data products) of the manuscript to enhance its visibility and encourage its use within the scientific community. 

The second is a hybrid data product that corrects long-term (1980–2019), high-resolution (~0.083° or 1/12°) modeled surface pCO2 for the Indian Ocean region (as a part of RECCAPv2) using cruise-based observations and an XGB algorithm. This product, available at https://www.nature.com/articles/s41597-025-04914-z (Ghoshal et al., 2025), falls under Section 3.1.6 (i.e., Model-based and hybrid data products and analysis) of this manuscript. 

In this study, a machine learning (ML) approach is employed to correct biases in surface pCO2 simulations generated by the INCOIS-BIO-ROMS model (pCO2model) over the period 1980–2019. The ML model is trained using the differences between observed (pCO2obs) and modeled pCO2 to estimate the spatio-temporal deviations (pCO2obs − pCO2model). These interannually and climatologically varying deviations are then added back to the original model output, resulting in two improved data products: pCIBR_Int and pCIBR_Clim. 

Evaluation against independent datasets, including moored observations (BOBOA), the gridded SOCAT product, and other ML-based pCO2 products (such as CMEMS-LSCE-FFNN and OceanSODA), demonstrates a significant improvement of approximately 40% ± 3.31% in RMSE compared to the original model. These corrected pCO2 products are expected to improve the accuracy of air–sea CO₂ flux estimates across the Indian Ocean from 1980 to 2019, helping to better identify key source and sink regions and enhancing our understanding of the Indian Ocean’s contribution to the global carbon budget. 

Further, in Section 3.1.6 (i.e., Model-based and hybrid data products and analysis), you may also consider including the model-based dataset and analysis of ocean acidification in the Indian Ocean from 1980 to 2019, as presented by Chakraborty et al. (2024). The paper is available at: https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2024GB008139. This study provides a comprehensive assessment of ocean acidification trends across the Indian Ocean and its sub-regions, utilizing outputs from a numerical model, an offline biogeochemical (BGC) model, and two machine learning-based products. Overall, the research consolidates the current state of knowledge on Indian Ocean acidification by integrating available field observations, reconstructed datasets, and model simulations. 

References: 

Joshi, A. P., Ghoshal, P. K., Chakraborty, K., & Sarma, V. V. S. S. (2024). Sea-surface p CO2 maps for the Bay of Bengal based on advanced machine learning algorithms. Scientific Data, 11(1), 384. 

Ghoshal, P. K., Joshi, A. P., & Chakraborty, K. (2025). An improved long-term high-resolution surface p CO2 data product for the Indian Ocean using machine learning. Scientific Data, 12(1), 577. 

Chakraborty, K., Joshi, A. P., Ghoshal, P. K., Baduru, B., Valsala, V., Sarma, V. V. S. S., Metzl, N., Gehlen, M., Chevallier, F., & Lo Monaco, C. (2024). Indian Ocean acidification and its driving mechanisms over the last four decades (1980–2019). Global Biogeochemical Cycles, 38(9), e2024GB008139. https://doi.org/10.1029/2024GB008139.