SINGLE-VARIABLE LOGISTIC REGRESSION MODELS OF STAG-HEADED APPEARANCE USING TREE RING DATA FOR LARIX GMELINII FROM CRYOLITOZONE IN CENTRAL EVENKIA

• Main Authors: Kachaev, A. V., Mashukov, D. А., V. Е. Benkova
• Format: Report
• Language: Russian
• Published: Zenodo 2021
• Subjects: logistic regression; stag-headed; radial growth; Central Siberia; Central Evenkia; Larix gmelinii Rupr. Rupr.; permafrost; Siberia
• Online Access: https://dx.doi.org/10.5281/zenodo.4521857; https://zenodo.org/record/4521857

Annual radial increment of trees is an integral descriptor of both biotic and abiotic processes in forest ecosystems. One of the approaches in tree mortality study is to build logistic regression models using tree ring data. Logistic variables are of different forms depending on the statistics chosen (e.g. mean, median, growth trend) and on the way they are counted in “windows” for the past N (5, 10, …, 40) years. We expanded the “window” system beginning from an arbitrary odd number between the ages of five to p (forty or fifty years of age), due to which expansion logistic variables and models analyzed increased in number. We found Larix gmelinii radial increment dynamics to differ between stag-headed and healthy individuals growing on a north-facing slope in permafrost conditions of central Evenkia (64°19´23˝ N, 100°13´28˝ E). The logistic regression models were built based on a single variable with the help of R programming language.