Hydroxyl radical (OH) is the most important oxidant in the troposphere that governs the oxidation power, while its spatial-temporal patterns are yet to be clear at present. The authors of this study developed an integrated data- and model-driven approach to predict the convoluted response of TOH to its five proxies, which include NO2, HCHO, H2O, Trop O3, and Strat O3. They investigated the trends and drivers of global, hemispheric, and regional OH from 2005 to 2019. Overall, this study has provided interesting results, which will help deepen our understanding of the changes in global tropospheric OH over the past decades. I have several concerns about the method that could impact the robustness of the results.

1) Is there any method to evaluate the response of OH to different input parametrizations that were calculated by Eq. (5)? These semi-normalized sensitivities laid the foundation for understanding the impacts from different proxies/drivers on tropospheric OH. I would like to see some aspects of “evaluation” or at least a comparison with previous other studies.

2) The input parameters were perturbed by the scaling factors of 1.1 and 0.9 in the ECCOH offline framework. What about the impacts of these assumed scaling factors? NOx and VOCs emissions might change by over 10% between 2005 and 2019, in both developed countries with advanced air pollution control and developing countries with rapid economic growth.

3) I feel that the trends in OH present in this study are substantially impacted by the trends of NO2 and HCHO, which were constrained by OMI. What about the impacts of uncertainties in the trends of OMI observations on the OH estimates? Especially for HCHO, I have seen some papers showing that there were systematic errors in the global and latitudinal trends.

4) Lines 498-500, the authors said “The correction factors, however, worsen the trends over the southeast US and Canada. This is essentially due to the use of the fractional errors in the a priori making the OMI corrections more impactful (i.e., higher Kalman gain) in summertime than in wintertime.” I did not understand what these sentences meant. Why did more impactful OMI corrections worsen the trends over the southeast US and Canada?