Melting snowfields are dominated by closely related green algae. Although microscopy–based classifi- cation are evaluable distinction tools, they can be challenging and may not reveal the diversity. High–throughput sequencing (HTS) allows for a comprehensive community evaluation but has been rarely used in such ecosystems. We found that assigning taxonomy to DNA sequences strongly depends on the quality of the reference databases. Furthermore, for an accurate identification, a combination of manual inspection of automated assignments, and oligotyping of the abundant 18S OTUs and ITS2 secondary structure analyses were needed. The use of one marker can be misleading because of low variability (18S) or the scarcity of references (ITS2). Our evaluation reveals that HTS outputs need to be thoroughly checked when the organisms are poorly represented in databases. We recommend an optimized workflow including consistent sampling, a two–molecular marker approach, light microscopy–based guidance, generation of appropriate reference sequences and a final manual verification of taxonomic assignments as a best approach for accurate diversity analyses.