Federated Learning (FL) enables collaborative model training across decentralized devices while preserving data privacy. Yet, its reliability in real-world, evolving environments is challenged by concept drift the continuous change of data distributions over time. This has led to the emergence of Federated Drift-Aware Learning (FDAL), which seeks to build adaptive and resilient FL systems. Unlike prior surveys that discuss individual challenges, this paper presents a structured synthesis and a taxonomy of FDAL’s research landscape. We categorize current limitations into four key areas: Data Characteristics and Adaptation, Knowledge Management and Retention, Security, Privacy, and Ethical Concerns, and Algorithmic and Computational Efficiency and highlight their interdependencies. Building on this analysis, we propose seven targeted research questions that outline a clear roadmap for advancing the field. This position paper contributes a fresh conceptual framework and a forward-looking research agenda, offering the community an integrated foundation for future FDAL development.