Artificial intelligence helps scientists study tundra shrubbery

Fig. 1. Young shrubby alder growing near closed shrub communities. Although shrub expansion creates a complex canopy structure, it usually results in a sharp decline in local species diversity. The herbaceous plants, dwarf shrubs, lichens, and mosses visible beneath this young shrub will likely disappear as the emerging shrub canopy becomes taller and denser.

An international team of researchers from the Severtsov Institute of Ecology and Evolution of the Russian Academy of Sciences (IEE RAS), HSE, the University of Virginia (UVA), and Alaska Biological Research applied artificial intelligence to study the distribution of shrub alder (Alnus alnobetula fruticosa) at the northern timberline in Siberia. The paper presents an innovative method for analyzing satellite images using convolutional neural networks. The technique opens up new possibilities for monitoring the effects of climate change in the circumpolar region, combining high accuracy with unprecedented speed of analysis.

The study covered three key areas in the Siberian Subarctic. The neural network analyzed almost a million 12x12 m plots based on high-resolution satellite images taken at intervals of 10-15 years. The results showed that the rate of shrub formation varies from 2.4% to 26.1% per decade depending on local conditions.

The main factors for shrub formation are heat supply and drainage: shrubs often occupy well-drained, warmed areas, as well as those located near existing shrub communities. The first stages of development of a new territory occur on flat, somewhat wetter surfaces, while closed shrub alder forests formed on warm and dry surfaces.

"The new method of data processing is hundreds of times faster than traditional approaches with high accuracy. The use of a neural network that classifies landscapes into four categories (from a complete absence of shrubs to closed shrub communities) made it possible to obtain a more detailed picture compared to simplified binary models," said Ksenia Ermokhina, PhD in Geographical Sciences, Senior Researcher at the IEE RAS.

Changes in vegetation associated with the spread of shrub alder have serious ecological consequences. Shrubs affect surface albedo, snow cover properties, and soil temperature, which can lead to an increase in the greenhouse effect in the region.

The work was published in the Journal of Ecology: Landscape patterns of shrubification in the Siberian Low Arctic, Anna Derkacheva, Gerald V. Frost, Howard E. Epstein, Ksenia Ermokhina, 02.08.2025.

Related materials:

RAS: "Artificial intelligence helps scientists study tundra shrubbery"