Typification of natural seasonal dynamics of vegetation to reveal impact of land surface change on environment (by satellite data)

Abstract

Abstract Deep insight into types of vegetation variability provided by AVHRR space scanner images of vegetation index spatial distribution helps reveal impact of land surface changes on environment. The Institute of Computational Modeling SB RAS has developed nonparametric algorithms of automatic to classify and recognize patterns of these images which helped to reveal: (1) major variability types (generally connected); (2) areas belonging to small classes, which can be used to reveal deviations from “normal” (e.g., forest fires, etc.); (3) deviation from a certain type of dynamics indicative of changes in condition of plants, which can be used to diagnose pathology at early stages; (4) impact of economical activities on vegetation in Norilsk area. The authors provide biological interpretation of the satellite data. Computer-animated dynamics and color maps are presented. Nonparametric algorithms of an automatic classification and pattern recognition were provided by the Institute of Computational Modeling SB RAS.