Are we missing the tail (and the tale) of residence time distributions in watersheds?

Residence times provide vital information on hydrological, geochemical, and ecological processes in watersheds. The common perception is that mean residence times in watersheds are very short, on the order of days to years. However, there is growing concern that longer residence times of centuries to millennia are not being captured by traditional surface water age-dating methods. We hypothesize that if mean residence times are biased short, then weathering rates calculated from mean residence times will be forced unrealistically high to match observed solute concentrations. We test this hypothesis by calculating weathering rates from springs based upon residence times estimated using three different age-dating methods. Observed solute concentrations require unrealistically large weathering rates if typical short residence times are employed as compared to rates derived from longer residence times. Residence time distributions in watersheds have substantially longer tails than previously thought, with implications for age-dating methods and their interpretation to infer process behavior. Key Points Watershed residence times are substantially longer than commonly thought Very short residence times were not supported by geochemical kinetics Geochemical kinetics provide unique test of physical meaning of residence times