The present work aims to utilise a scale-adaptive extension of the baseline Spalart-Allmaras turbulence model to propose a new way for turbulence model uncertainty quantification. During previous work, the extension waswas implemented into the DLR TAU code and it was proven that the extension of the baseline SA-model can be reliably used to control the size of resolved scales and consequently the ratio between modelled and resolved flow physics. A bi-modal solution behaviour was observed, where for large values of CSAS the model produced URANS-like statistical results, while moderate and small values produce significantly improved results over the baseline URANS calculations. The present work aims to build upon the findings to present a model uncertainty framework that uses the relationship between the ratio of the resolved to total turbulent kinetic energy and the solution variables for different values of CSAS to estimate the modelling error of the underlying RANS-model. Computations are carried out for a generic triple delta-wing configuration [4] under transonic side-slip conditionsat non-stall and stall angles of attack. The evolution of the aerodynamic force and moment coefficients is investigated starting from URANS with increasing fidelity computations to quantify the model error.In contrast to most existing model uncertainty quantification approaches, which typically consider sen-sitivities within their framework model, the proposed way heads towards a comprehensive model uncertainty formulation by introducing more physics into the flow-field solution.