Bluecat: A Local Uncertainty Estimator for Deterministic Simulations and Predictions

We present a new method for simulating and predicting hydrologic variables with uncertainty assessment provide example applications to river flows. The is identified the acronym “Bluecat” based on use of deterministic model which subsequently converted stochastic formulation. latter provides an adjustment statistical basis prediction along its confidence limits. distinguishing features proposed approach are ability infer probability distribution without requiring strong hypotheses characterization error (e.g., normality, homoscedasticity), transparent intuitive observations. Bluecat makes rigorous theory estimate predictand conditioned by output, inferring conditional statistics Therefore bridges gaps between (possibly physically based, or deep learning-based) models, as well data innovative user oriented approach. two examples application case studies Arno at Subbiano Sieve Fornacina. results confirm technical soundness. open software working in R environment, help facilities detailed instructions reproduce presented here.