The loss of biodiversity and the associated decline of ecosystem services vital for sustaining human life demand a comprehensive monitoring of plant biodiversity. Measuring biodiversity in the field on large areas generates issues like the need of a robust sampling design, the high demand on human and monetary resources and different biases introduced by humans and environmental conditions. These circumstances have recently triggered an extended use of remote sensing data to quantify biodiversity in a cost- and time-efficient way. Remotely sensed datasets represent the Earth surface at a certain point in time. Yet, it is not well studied what the use of a single dataset in time implies for biodiversity estimates. The functional dimension of biodiversity, expressed through functional traits within or between species, varies according to the phenological cycle. Further in grasslands, mowing and grazing events lead to temporal variations in the remotely sensed diversity. We provide an approach in which we integrate the temporal dimension in the quantification of biodiversity from space. Functional diversity is partitioned into a spatial and a temporal component. In particular, Sentinel-2 satellite datasets are well suited for this purpose, providing a complete landscape picture with high revisit time. In our study case, the incorporation of the temporal dimension and the interaction between spatial and temporal diversity by employing multiple datasets improves the retrieval of functional diversity in differently managed alpine grasslands. In comparison to the use of a single dataset, our approach provides more reliable recommendations for conservation and restoration decision-making on a regional scale.