Evaluation of gridded and satellite precipitation product over hilly terrain of Meghalaya (India) for water management

For accurate assessment and forecasting of rainfall, a network of rain gauge stations is required with decent density. However, the availability of rain gauges particularly in the Northeastern India is a serious concern. To overcome the limitation of observed data, SPPs are becoming convenient as it has an easy accessibility and good spatial resolution. The present study aims to evaluate the applicability of SPPs and gridded data as against observed data at different temporal scale (viz., daily, monthly and annual) in the state of Meghalaya in NE India. The comparison was made on the basis of continuous and categorical matrices. These matrices give quantitative and qualitative comparisons of the two datasets. The study indicates both gridded and SPP have a very high error in Sohra and Mawsynram stations in all the cases (daily, monthly, and annually) and generate a negative correlation. However, a good correlation was observed for IMD with observed data than IMERG at monthly scale with r range from - 0.414 at Sohra to 0.803 at Sasangre station, 134.80-489.51 mm/month (RMSE), 76.54-454.97 mm/month (MAE), - 100.99 to 59.71 (BIAS). Similarly, as per the categorical indices, both gridded and SPP showed errors in detecting capabilities. Both Sohra and Mawsynram stations are among the station that receives the extreme rainfall globally. Thus, the errors can be attributed to the terrain conditions. However, IMD detected the rainfall better as compared to IMERG datasets.