Environmental flow (EF) is a widely used concept in river basin management and river ecology. There are more than 200 methods to estimate the flow that should be left to the rivers to ensure a sustainable resource management, and it is almost impossible to standardize a methodology that could be applied for each stream everywhere in the world.
Although some global exercises have been performed (IWMI for instance), in fact, the EF is extremely site and season specific: it depends on the natural conditions of the river, its hydrological, geo-morphological, and biological features, and its natural ecosystem. That for, the use of specific methodologies makes sense in the management of a specific river basin or a portion of it, but what about global hydrological analyses?
I am setting up a global gridded hydrological analysis and, while accounting for water uses, I would like to set up a threshold under which the water withdrawal is not allowed. I am thinking of using a percentile of the historical modeled monthly time series of discharge spanning a period of about 30 years. This would allow to set the threshold to a value for each month that is a percentile of the historical undisturbed (natural, with no fragmentation or withdrawal) discharge, for each of the grid-cells. Is this an acceptable approximation?