Causal reasoning: Towards dynamic predictive models for runoff temporal behavior of high dependence rivers

Abstract

Nowadays, a noteworthy temporal alteration of traditional hydrological patterns is being observed, producing a higher variability and more unpredictable extreme events worldwide. This is largely due to global warming, which is generating a growing uncertainty over water system behavior, especially river runoff. Understanding these modifications is a crucial and not trivial challenge that requires new analytical strategies like Causality, addressed by Causal Reasoning. Through Causality over runoff series, the hydrological memory and its logical time-dependency structure have been dynamically/stochastically discovered and characterized. This is done in terms of the runoff dependence strength over time. This has allowed determining and quantifying two opposite temporal-fractions within runoff: Temporally Conditioned/Non-conditioned Runoff (TCR/TNCR). Finally, a successful predictive model is proposed and applied to an unregulated stretch, Mijares river catchment (Jucar river basin, Spain), with a very high time-dependency behavior. This research may have important implications over the knowledge of historical rivers' behavior and their adaptation. Furthermore, it lays the foundations for reaching an optimum reservoir dimensioning through the building of predictive models of runoff behavior. Regarding reservoir capacity, this research would imply substantial economic/environmental savings. Also, a more sustainable management of river basins through more reliable control reservoirs' operation is expected to be achieved.