Liverwort altitudinal patterns in East Kamchatka: dividing the indivisible

Small cryptogamic plants, including liverworts, are confined in their occurrence to microhabitats and may be weakly associated with communities formed by dominant vascular plant species. This complicates the understanding of patterns in their elevational distribution in mountainous areas - a task that becomes doubly challenging under conditions of poorly expressed altitudinal zonation.

Based on the concept of a Gaussian distribution model, modified in the present study to account for constraints imposed by the characteristics of the material, potential distribution curves and occurrence frequencies were calculated for each species using collections made in the study area in East Kamchatka.

Four elevational groups of liverworts were identified based on the occurrence-record adaptation of the range-density maximum likelihood estimation framework. The transition from one elevational group to the next is characterized by a pronounced trough in the positions of the elevational core centers of species and clearly defines each new group identified using the method applied herein. Species lists obtained using the abovementioned method were compared with taxonomic spectra for five elevational belts empirically recognized based on vascular plant characteristics. The groups identified under each approach partially coincide. However, the pyroclastic-field flora cannot be recognized as a distinct elevational group.

The method allows the identification of elevational groups under conditions where gradual changes and deep interpenetration of adjacent communities along the elevational profile preclude the recognition of distinct vegetation belts. Even where elevational vegetation complexes can be recognized, the method applied here appears preferable to delimiting elevational groups based on dominant communities, since the distribution of liverworts is tied more closely to substrates than to communities formed by dominant vascular plant species. The applied method may be useful for incomplete data on altitudinal distribution, as the expected distribution is modeled using range-density maximum likelihood estimation.