Understanding how different taxa respond to abiotic characteristics of the environment is of key interest for understanding the assembly of communities. Yet, whether eDNA data will suffice to accurately capture environmental imprints has been the topic of some debate. In this study, we characterised patterns of species occurrences and co-occurrences in Zackenberg in northeast Greenland using environmental DNA. To explore the potential for extracting ecological signals from eDNA data alone, we compared two approaches (visual vegetation surveys and soil eDNA metabarcoding) to describing plant communities and their responses to abiotic conditions. We then examined plant associations with microbes using a joint species distribution model. We found that most (68%) of plant genera were detectable by both vegetation surveys and eDNA signatures. Species-specific occurrence data revealed how plants, bacteria and fungi responded to their abiotic environment – with plants, bacteria and fungi all responding similarly to soil moisture. Nonetheless, a large proportion of fungi decreased in occurrences with increasing soil temperature. Regarding biotic associations, the nature and proportion of the plant-microbe associations detected were consistent between plant data identified via vegetation surveys and eDNA. Of pairs of plants and microbe genera showing statistically supported associations (while accounting for joint responses to the environment), plants and bacteria mainly showed negative associations, whereas plants and fungi mainly showed positive associations. Ample ecological signals detected by both vegetation surveys and by eDNA-based methods and a general correspondence in biotic associations inferred by both methods, suggested that purely eDNA-based approaches constitute a promising and easily applicable tool for studying plant-soil microbial associations in the Arctic and elsewhere.