Assessment of droughts and their linkage to environmental flow conditions over a large Indian river basin

A robust characterization and risk assessment of meteorological droughts is the need of the hour considering its pervasiveness and consequences; however, their precise physical quantification is a difficult geophysical endeavor. This becomes a serious issue for India, having 18% of the world’s population and 4% of global freshwater, out of which 83% is used in agriculture. In this study, a detailed spatiotemporal assessment of the meteorological droughts characterized by standardized precipitation index (SPI) at annual scale is carried out over the Narmada Basin, India using the monthly rainfall data from 24 stations for 63 years (1951- 2013). The entire duration was divided into two epochs of 31 years (i.e. 1951-1981 and 1982-2012) for a comparative assessment of drought characteristics. The non- parametric Mann- Kendall (MK) test is applied to investigate the trend of droughts. Further, to predict the environmental Flow (EF) conditions from rainfall data only, the linkage of SPI with the average annual flow (%AAF) is examined over four sub-catchments (Mohegaon, Hridaynagar, Manot, and Sher) of the basin. The results reveal that the Narmada basin is prone to droughts with a frequency of once in 3 to 5 years. The frequency and severity of droughts have significantly increased in 1982-2012 as compared to 1951-1981. The severity of recent droughts shows a more widespread aerial extent in the region. The MK test results indicate an increasing trend in the droughts over most of the stations. An exquisite agreement between SPI and %AAF (used to describe the EF condition) is observed with R² ranging from 0.757 to 0.988, which shows that coupling SPI with %AAF can be effective for ungauged catchments. This study suggests that appropriate measures must be taken for better management of the water resources in the basin, and also for mitigation droughts, considering the increased risk of the severe drought events in recent decades.