Approximate Bayesian Computation for Forecasting in Hydrological models

Abstract

Approximate Bayesian Computation (ABC) is a statistical tool for handling parameter inference in a range of challenging statistical problems, mostly characterized by an intractable likelihood function. In this paper, we focus on the application of ABC to hydrological models, not as a tool for parametric inference, but as a mechanism for generating probabilistic forecasts. This mechanism is referred as Approximate Bayesian Forecasting (ABF). The abcd water balance model is applied to a case study on Aipe river basin in Columbia to demonstrate the applicability of ABF. The predictivity of the ABF is compared with the predictivity of the MCMC algorithm. The results show that the ABF method as similar performance as the MCMC algorithm in terms of forecasting. Despite the latter is a very flexible tool and it usually gives better parameter estimates it needs a tractable likelihood