Precipitation extremes and anomalies of the Indian Sundarban 1984-2018

Abstract

Gridded precipitation data products of 0.5º × 0.5º spatial resolution were analysed to understand the climatic variability in a spatial and temporal context. Data reliability of processed gridded data products were examined in the absence of gauge station data observations in the study area. However, the implementations of comparative analysis of the spatial and temporal data products in this study area are missing. The NASA Power Data (NPD) and Climate Research Unit (CRU TS 4.03) Data were scrutinized from 1984-2018. The data products were selected, compared, and interpreted grid wise. Annual and monsoonal precipitation pattern was also studied. Data variability has been analyzed using the Coefficient of Variation (CV), Anomaly, and Precipitation Concentration Index (PCI). The statistical analysis of R², MAE, RMSE, MAPE and BIAS was performed to quantify the error and differences. Considering the independent grid point, the MAPE and BIAS indicate that only grid 4 performed better than the rest with 12.7% and 17%, respectively. The results regarding the data products illustrate significant differences both in averaged and grid wise context. The NPD shows an increasing trend, whereas CRU represents a decreasing trend from the year 1984-2018. Before the implementation of any processed secondary gridded data products in complex terrain, the critical evaluation and quantification of the magnitude of error is a prerequisite, like the Sundarbans, where the gauge stationed data is unavailable.