АКТУАЛИЗАЦИЯ ЛАНДШАФТНОЙ КАРТЫ ПИНЕЖСКОГО ЗАПОВЕДНИКА МЕТОДАМИ МНОГОМЕРНОГО АНАЛИЗА

• Main Author: Попов Сергей Юрьевич
• Format: Text
• Language: unknown
• Published: Фонд поддержки и развития заповедных территорий «Медвежья земля» 2016
• Subjects: ландшафтная карта; ГИС-технологии; Landsat; цифровая модель рельефа; дискриминантный анализ; Lambda; Pinega; Arkhangelsk; taiga
• Online Access: http://cyberleninka.ru/article/n/aktualizatsiya-landshaftnoy-karty-pinezhskogo-zapovednika-metodami-mnogomernogo-analiza; http://cyberleninka.ru/article_covers/16882899.png

В статье рассматривается цикл работ по актуализации ландшафтной карты Пинежского заповедника. Для актуализации пространственной информации использовался дискриминантный анализ, точность которого составила 90.2%. Приводится подробный разбор качества анализа. При актуализации удалось выделить два новых ландшафта. Актуализированная ландшафтная карта Пинежского заповедника имеет более точные границы и может использоваться для дальнейших исследований на современном уровне. The article discusses the cycle of updating the landscape map of the Pinega State Reserve. This reserve is located in the Arkhangelsk District, in the northern taiga subzone (64°35N, 42°58E) in Russia. A landscape map of the reserve made in 1999 was established on the basis of only the topographical maps and is very schematic. The modem technique of the combined spatial data stepwise discriminant analysis (SDA) of Landsat and DEM data was used aiming to update the map of landscapes of the Pinega State Reserve. The stepwise criterion of lambda values have shown a high level of SDA reliability (the quality of classification is 90.2%). The independent variables for PDA were the brightness characteristics of Landsat channels, the spectral indices and the characteristics of the terrain. A numbers of 73 independent variables have been used in the analysis. Areas of the old landscape map were used as a predictor. All operations with spatial and statistical data are made in the geo-information packages ArcGIS, ERDAS IMAGINE, IBM SPSS, ENVI, Statictica. A detailed analysis of the quality of SDA is provided. Two new landscape types have been revealed as a result of using SDA; moreover, the updated map of landscapes of the Pinega State Reserve has more accurate boundaries now. Obtained on the basis of multivariate analysis the landscape map of the Pinega State Reserve is statistically significant. Landscape squares are calculated.