How old is the water in the Yellow River Source Area? A multiscale analysis using transit time distributions

Abstract

Estimating water yield is a crucial aspect of evaluating conservation strategies and ensuring sustainable development of watershed water resources. The determination of the transit time distribution (TTD) using stable water isotopes has become essential for characterizing the storage and release processes of water in watersheds. In this study, we estimated TTD using the convolution integral method during the thawing period of five permafrost-influenced, nested basins (20000∼120000 km²) in the Yellow River Source Area (SAYR). To account for the effect of the thawing permafrost layer, an additional source contributing to runoff was added to the convolution integral approach. We derived the mean transit time (MTT) and the fraction of water younger than threshold age (hereafter abbreviated as T.A, (F(t < T.A)) from the TTD. The result of Bayesian optimization showed that F(t < T.A) ranged from 0.12 to 0.64, and MTT ranged from 40 to 368 days. The F(t < T.A) and the MTT of the watersheds were mainly influenced by river channel topography and the water retention capacity of the watersheds which depended on the topographical and morphological characteristics. They were to a lesser degree affected by land use type, soil type, and the thermal stability of the frozen soil. The results in this study show that estimating watershed transit times could be improved by incorporating a convolution integral model with permafrost meltwater contribution during melting time.