Fig 1.
Interpreting Ellenberg Indicator Values.
Ellipses showing the multivariate normal distribution of vegetation types plotted along Ellenberg indicator values for light, soil moisture, pH and nutrient status. Agricultural includes rotational fields, old fields and lays. The colors of the ellipses correspond to characteristic sites depicted on the right.
Table 1.
Habitat types and geographic origins and their description.
Table 2.
Predicting environmental gradients of light, soil moisture, pH and nutrient status from variation in soil eDNA.
Table 3.
Predicting habitat types from variation in soil eDNA.
Fig 2.
Mock crime scene 1 and the provenance derived from soil eDNA.
Predicted location in environmental space (Ellenberg Indicator Values) based on the best linear models and GAM for light (top left). The cross marks the predicted EIV value and the length of the arms show the 95% confidence intervals. Blue dots show the actual EIVs at the sample site. The predicted probability of binary habitat types based on Quadratic Discriminant Analysis are shown as red lines on a box plot of the distribution of predicted values for each characteristic (middle part). A priori classification membership is indicated by asterisk of the 0/1 variable. The probabilities for heathland, coniferous and Atlantic are not shown. A list of plant species sequences, their frequency of sequences (FS) in the sample, and if applicable, the geographic region of their distribution in Denmark is listed bottom left. For evaluation, we provide a picture of the sample location (bottom right).
Fig 3.
Mock crime scene 2 and the provenance derived from soil eDNA.
Predicted location in environmental space (Ellenberg Indicator Values) based on the best linear models and GAM for light (top left). The cross marks the predicted EIV value and the length of the arms show the 95% confidence intervals. Blue dots show the actual EIVs at the sample site. The predicted probability of binary habitat types based on Quadratic Discriminant Analysis are shown as red lines on a box plot of the distribution of predicted values for each characteristic (middle part). A priori classification membership is indicated by asterisk of the 0/1 variable. The probabilities for heathland, coniferous and Atlantic are not shown. A list of plant species sequences, their frequency of sequences (FS) in the sample, and if applicable, the geographic region of their distribution in Denmark is listed bottom left. For evaluation, we provide a picture of the sample location (bottom right).