Model-to-model transformations constitute an important ingredient in model-driven engineering. As real world transformations are complex, systematic approaches are required to ensure their correctness. The ATLAS Transformation Language (ATL) is a mature transformation language which has been successfully applied in several areas. However, the executable nature of ATL is a barrier for the validation of transformations. In contrast, transformation models provide an integrated structural description of the source and target metamodels and the transformation between them. While not being executable, transformation models are well-suited for analysis and verification of transformation properties. In this paper, we discuss (a) how ATL transformations can be translated into equivalent transformation models and (b) illustrate how these surrogates can be employed to validate properties of the original transformation.