An accurate spatial and temporal representation of rainfall is essential for hydrological assessments and water resources management. Rainfall is monitored in India’s mountainous Western Ghats region via in-situ rainfall gauging stations maintained by the Indian Meteorological Department (IMD). However, the network is sparse, and significant periods of data are missing. Furthermore, the IMD gridded rainfall dataset is known to underestimate the depth of rainfall at the high altitudes within this region. In this study, rainfall estimated by the IMD grids and from remote sensing using the CHIRPS (0.25- and 0.05- degree), MSWEP and PERSIANN datasets are compared to the IMD in-situ gauged rainfall within the Western Ghats using a point-to-pixel analysis.

The GWAVA model is utilised to determine the effect of the selected rainfall input datasets on representing wider water resources. It was found that the average ensemble provided the best representation of the in-situ gauged and catchment rainfall and a better representation than the IMD grids. It remains critical for water resources management to ensure that in-situ rainfall gauging networks are maintained. In-situ data sources of high confidence remain important for the continuous development and ground-truthing of different rainfall datasets.