Predicting the Evolution of Epigeic Arthropods in a Vineyard Landscape Using Machine Learning

Abstract

The agricultural landscape, which includes viticulture and grape growing, significantly affects the spatial distribution and the development of biodiversity in the landscape. Therefore, in our study, we focused on the development of epigeic arthropods under the conditions of the vineyard landscape. We conducted the research over the three years 2021-2023 in 8 study areas representing 4 types of vineyard landscape use (Intensive vineyard, semiintensive vineyard, abandoned vineyard, meadow (original vineyard)). We collected epigeic arthropods using the pitfall traps method at monthly intervals. During the research, we recorded 56,403 individuals belonging to 22 taxa, eudominant representation was confirmed in the taxa Hymenoptera (40.59%), Coleoptera (18.67%), and Araneida (11.65%). Statistical analysis confirmed the highest number of species and individuals in the intensive vineyard and meadow (original vineyard). Using machine learning, we confirmed the same evolution of epigeic arthropods over time for all types of vineyard land use except semiintensive vineyard. We also predicted an increase in epigeic arthropods for the month of July 2025 for all types of land use. The results of our research provide new information on the development of epigeic arthropods in vineyard ecosystems, which is important for the maintenance of sustainable agriculture that supports the preservation of biodiversity.