In temporal knowledge graph completion (TKGC), translation, rotation, and scaling have proven effective in capturing temporal dynamics. To comprehensively model the diverse relationships in temporal knowledge graph embeddings (TKGE), it is crucial to incorporate multiple transformations. By employing translation, scaling, and rotation along the temporal dimension, evolving semantic information can be captured with greater precision. Biquaternions, which can seamlessly combine multiple geometric transformations—including scaling, translation, Euclidean rotation, and hyperbolic rotation—are particularly suitable for this purpose. In this work, we present TCompoundQ, a model built within the quaternion vector space that unifies translation, rotation, and scaling operations through biquaternions. Experimental results on five publicly available temporal datasets validate the effectiveness of TCompoundQ, showcasing its superior performance.