Over the past few decades, hybrid RANS/LES methods for modeling turbulence have become a popular compromise between the computationally cheap Reynolds Averaged Navier-Stokes (RANS) approach, and the reliable Large Eddy Simulation (LES). But they still face conceptual issues when transitioning from RANS to LES (or reversely), such as log-layer mismatch, modeled scale depletion, globally labeled as grey area issues. From a theoretical point of view, this issues may be related to the commutation error arising when the hybrid RANS/LES filter varies rapidly. Following the additive filter method, initially presented by Germano (Theoretical and Computational Fluid Dynamics, 17(4):225–231, 2004 ), we present a method aimed at estimating the aforementioned commutation errors.