An Artificial Intelligence in the Simulation of Tree-Ring Growth around the World
The work addresses a fundamental problem of forest reaction forecast to the climate change and increasing concentrations of greenhouse gases for the terrestrial ecosystems of the Earth.
The problem of tree-ring response to possible climate change is one of the most urgent problems of the modern forest ecology. Despite the large number of papers concerning tree-ring response to different environmental changes (temperature increasing, drought, etc.) there is no reliable answer how woody plants will respond to environment changes in different forest stands and various physiographic zones.
The main target is to produce a retrospective assessment and a short-term forecast of annual tree-ring productivity (seasonal cell production) of the major conifer plant species in terrestrial forest ecosystems around the World forced by climate and non-climatic factors. The analysis will be based on an Artificial Intelligence Platform “Global Tree-Ring Growth Evolution Neural Network (VS-GENN)” and available datasets developed for the European, Asian and South American dendrochronological test-polygons.
The Global Tree-Ring Growth Evolution Neural Network by itself is a parameterization procedure of the VS-model which combines three novel parallel process:
- Direct parameterization based on optimization evolutional IT-algorithm;
- Proxy parameterization based on a VS-metamodel (artificial neural network which operates as direct VS-model but can produce cell profiles much faster)
- Re-training of VS-metamodel to reduce a discrepancy between simulation tree-ring growth curves obtained by the direct and proxy parameterizations.
The work has no analogues in Europe. All stated tasks are novel and extremely important for understanding global processes undergoing in the forest ecosystems under observed climatic changes and other disturbances.
LE STUDIUM / Marie Skłodowska-Curie Research Fellowship
Pr Vladimir Shishov
FROM: Siberian Federal University - Mathematical Methods and IT Department, Krasnoyarsk - RU
IN RESIDENCE AT: BioForA, Centre Inra Val-de-Loire / ONF - FR