An automatic bioaerosol classification system is an attractive alternative to the standard visual identification and counting of pollen in the standard volumetric method of aerobiological monitoring. While various physical principles can be used for automatic measurement of the parameters of particles present in the air, the key problem becomes the development of a classification model based on these data. In particular, practical application of the models becomes challenging due to the large variability of particles present in real air compared to laboratory experiments in which models are usually trained. Instead of training the model on data obtained in laboratory conditions, we applied a clustering algorithm to fluorescence spectra collected during daily measurements in the 2024 season with the Rapid-E+ automatic detector installed at the monitoring station at Moscow State University. Comparison of the temporal distribution of particles in each cluster with the seasonal dynamics of eleven pollen types obtained from standard aerobiological monitoring with a volumetric trap at the same station at Moscow State University shows that some clusters (i.e., fluorescence spectra of specific shape and amplitude) demonstrate temporal patterns similar to pollen seasons. However, the fluorescence spectra alone are not sufficient for differentiation of individual pollen types, and they can only provide detection of larger groups of bioaerosols. Interestingly, the detected larger groups show more diverse seasonal patterns than those observed by volumetric monitoring at the station at Moscow State University. This result demonstrates that automatic detectors can provide more useful information on the content and seasonal distribution of bioaerosols compared to standard volumetric methods.