Range shift and climatic refugia for alpine lichens under climate changes

Mountain biodiversity is highly threatened by climate change, with many species facing habitat reduction and/or local extinction. Poikilohydric organisms like lichens are extremely sensitive to environmental conditions and changes. A comprehensive overview of the impact of climate change on the future distribution of lichens is still missing. Thus, we quantified the range shift, loss in suitable areas, and potential climatic refugia for more than 250 lichens, exploring these trends for each species and grouping them by their ecological needs and functional traits. We performed species distribution models to investigate the range shifts of 272 lichens under different climate change scenarios. Range shift was investigated by modelling gain and loss in suitability score along the elevational gradient, for each species separately, and grouping them based on growth forms and temperature affinities. Based on the "high suitability score stability" index, we developed maps of potential refuge areas.  We predicted a huge loss of habitat suitability over time, especially under the SSP5-5.8. Fruticose lichens will exhibit the highest loss, and cryophilous species will be the most impacted. Moreover, contrary to the assumption of an upward range shift, most species showed no significant relationship between altitude and an increase in habitat suitability, suggesting different redistribution patterns for lichens. In the same way, climatic refugia are not only identified at high elevations but also in deep valleys and cold exposure. The response of lichens to climate change is extremely heterogeneous, and an upper-range shift is less common than expected. Spatial redistribution is highly distinct among functional groups. The identified climatic refugia confirm this pattern, underlining the urgent need to improve the conservation effort, especially for fruticose and cryophilic species.

Methods
The lichen occurrences were derived from the following sources: Dolichens (https://italic.units.it/dolichens/), ITALIC 7.0 (https://italic.units.it/), GBIF (https://www.gbif.org/). We retained only occurrence data collected after 1975 and with a coordinate uncertainty of 500 m or less. To reduce spatial bias due to pseudo-replication of occurrences, we performed spatial thinning. Then, we retained only the species with at least 25 occurrence records, a limit that has been shown to provide reliable results in modeling studies. The resulting dataset includes 25 to 469 occurrence records of 272 infrageneric taxa, for a total of 23,521 occurrences. For each species, we generated 10 replicate sets of background points, in a 10:1 ratio with the number of presence points, with an exclusion buffer of 500 m from the presence point.