Technical note: RAT – a robustness assessment test for calibrated and uncalibrated hydrological models

Abstract. Prior to their use under future changing climate conditions, all hydrological models should be thoroughly evaluated regarding temporal transferability (application in different time periods) and extrapolation capacity beyond the range of known past conditions). This note presents a straightforward evaluation framework aimed at detecting potential undesirable dependencies models: robustness assessment test (RAT). Although it is conceptually inspired by classic differential split-sample Klemeš (1986), RAT advantage being applicable types models, they calibrated or not (i.e. regionalized physically based). In this note, we present RAT, illustrate its application on set 21 catchments, verify applicability hypotheses compare previously published tests. Results show that an efficient approach, passing successfully can considered prerequisite for any model used change impact studies.