Geospatial modelling of the widespread broadleaved forest syntaxa of order Carpinetalia betuli P. Fukarek 1968 in the North-Western Caucasus

This study was the first to carry out geospatial modelling of broadleaved forest syntaxa using machine learning methods for the territory of the Caucasus. In total, the article used data from 278 vegetation relevés obtained by the authors in the broadleaved forests of the North-Western Caucasus in 2014–2019. Modelling was performed for the 6 most common syntaxa of oak-hornbeam forests of order Carpinetalia betuli P. Fukarek 1968: ass. Aro maculati–Carpinetum betuli Shevchenko et Braslavskaya, 2021, ass. Tamo communis–Carpinetum betuli Shevchenko et Braslavskaya, 2021, its variants typica, Staphylea colchica, and Festuca drymeja, and the community Abies nordmanniana–Carpinus betulus. The area of all 6 studied syntaxa was determined and distribution maps were created. Spatial differentiation (that is, no overlapping areas) was revealed in the most widespread (in terms of area) ass. Tamo communis–Carpinetum betuli and Aro maculati–Carpinetum betuli, confirming the specificity of the bioclimatic confinement of these syntaxa. It was found that the most important bioclimatic predictors in model development of all 6 described syntaxa were indicators of Precipitation Seasonality (bio15) and Precipitation of Warmest Quarter (bio18). It was established that over the medium term, if the trend of climate change and deforestation persists, the square of the potential area of ass. Aro maculati–Carpinetum betuli in the low-mountain and especially in the foothill zone of the North-Western Caucasus will reduce, whereas the square of ass. Tamo communis–Carpinetum betuli and all its described variants will increase. In the low- and mid-mountain zones, the area of community Abies nordmanniana-Carpinus betulus will increase.