All reviewers recognise that the scope of this manuscript represents an interesting and relevant topic, relating to the uncertainty in precipitation data products in a region of relatively poor instrumental data. However, all reviewers identify important inadequacies in the current version, which can be summarised as a :
• Lack of in-depth spatio-temporal comparison and analysis of the absolute precipitation values within the datasets;
• Lack of rationale for the selection and length (5 years) of the inter-comparison period, and whether any differences found are meaningful in light of inter-annual variability and large-scale meteorological forcings;
• Lack of consideration of the implications of the different spatial scales (grid sizes) of the data products
• Lack of any benchmarking to determine which is the ‘best’ dataset for use in hydrological modelling and other types of studies, particularly given the uncertainty in deriving actual precipitation amounts from instruments that do not adequately capture spatial (and vertical) heterogeneity in these data sparse regions

The authors’ response to the reviewers lacks specificity but suggests that they will address the concerns raised - for the paper to be published, the identified weaknesses will need to be comprehensively addressed. They identify that they do not have access to the IMD 0.25o dataset for use as a benchmark, but propose to utilise observed river flow and MODIS ET data. This is an interesting approach, although its utility will depend upon the spatial scale of the river basins with available data (i.e. whether their extent leads to excessive spatial averaging) and the importance of agricultural irrigation in the river basins (and therefore how different gauged flows are from naturalised flows)

The paper address an important issue faced by hydro-meteorologists and has the potential to have significant utility for researchers and practitioners in the region. However, Reviewer #4 has identified that, whilst the revised manuscript has addressed some of the original concerns raised, there remain important problems in the strength and sufficiency of the methodology and of the analysis of the results. These are clearly articulated in the in-depth comments in the uploaded referee report for the authors to further improve the manuscript. In particular, there remains a lack of robust benchmarking to determine which is the “best” dataset for use in hydrological modelling and the reviewer has suggested using freely available WMO observations of daily/monthly precipitation to overcome this major issue.

The authors have comprehensively addressed the requested revisions which have greatly increased the quality and utility of the paper. The paper is accepted