River ecosystem resilience risk index: A tool to quantitatively characterize resilience and critical transitions in human-impacted large rivers.

Riverine ecosystems can have tipping points at which the system shifts abruptly to alternate states, although quantitative characterization is extremely difficult. Here we show, through critical analysis of two different reach scale (25 m and 50 m) studies conducted downstream of two point sources, two tributaries (main stem and confluences) and a 630 km segment of the Ganga River, that human-driven benthic hypoxia/anoxia generates positive feedbacks that propels the system towards a contrasting state. Considering three positive feedbacks-denitrification, sediment-P- and metal-release as level determinants and extracellular enzymes (β-D-glucosidase, protease, alkaline phosphatase and FDAase) as response determinants, we constructed a 'river ecosystem resilience risk index (RERRI)' to quantitatively characterize tipping points in large rivers. The dynamic fit intersect models indicated that the RERRI<4 represents a normal state, 4-18 a transition where recovery is possible, and >18 an overstepped condition where recovery is not possible. The resilience risk index, developed for the first time for a lotic ecosystem, can be a useful tool for understanding the tipping points and for adaptive and transformative management of large rivers.