Inveterarens underbara resa genom Jämtland - ett test av träffsäkerheten i artutbredningsmodell för norna (Calypso bulbosa) med fokus på Jämtlands län

Sammanfattning: As a member state of the European union, Sweden is obligated to report the status of the species and habitats protected under the Habitats Directive every sixth year. Calypso bulbosa is one such species. Currently, there is suspicion that the true Swedish population size might have been heavily underestimated, as the habitat the species inhabits is relatively common. To address this uncertainty, Stephan et al. (2022) from SLU Artdatabanken produced a species distribution model (SDM) relating the presence/pseudo absence of C. bulbosa to a series of selected environmental factors. The model was then used to map potential growth sites as continuous probabilities across northern Sweden on a hectare level. In this report I test the accuracy of the above-mentioned model and provide data to be used in future improvements of this model. I visited 13 areas across Jämtland county, collecting presence/absence data from hectare squares with varying probabilities of presence according to the species distribution model. Logistic regression was then used on four data sets to determine if the probabilities of the model could explain the presence/absence observed in the field. I also recorded presence/absence of eight other plant species to determine if their presence/absence were associated with the presence/absence of C. bulbosa.  The results showed a significant positive relationship (p = 0,0167) between the probabilities of the SDM and the presence of C. bulbosa recorded in the field for one of the data sets. This data set included data from a simultaneous study in the Lapland part of Norrbotten county but excluded locations with previous recordings of the species. However, the probabilities only explained 1% of the variation in the field data. There were indeed also significant associations between the presence of four out of eight suggested associated species and that of C. bulbosa. These species were, in order of declining association: Goodyera repens, Maianthemum bifolium, Daphne mezereum and Gymnocarpium dryopteris.  Evidently, more factors than those used in the SDM affect the distribution of C. bulbosa. Nevertheless, the results show that the model could be made more useful by using the newly collected data to make future improvements.