A TECHNIQUE OF SPATIO-TEMPORAL ANALYSIS OF DARKNEEDLE STANDS DESICCATION BASED ON LANDSAT REMOTE SENSING DATA

The goal of this research was to develop a cost-effective technique to analyze spatio-temporal dynamics of darkneedle stands desiccation. The developed technique allows estimating of spatio-temporal dynamics of darkneedle stands desiccation based on remote sensing data from Landsat satellites regarding orography and climate trends. Advantages of the technique are (1) using of freely available Landsat data, digital elevation model and climate data; and (2) it is based on the maximum likelihood supervised classification method realized in the most of software products. There are six main steps in the technique: (1) preliminary data preparation and analysis; (2) generation of classification map of darkneedle stands for the period prior to decline of trees; (3) masking of time series of Landsat data based on the classification map of darkneedle stands for the period prior to decline of trees; (4) generation of classification maps of desiccated stands; (5) GIS-analysis of relationships between spatio-temporal dynamics of desiccation of trees and orography; (6) statistical analysis of relationships between spatio-temporal dynamics of desiccation of trees and climate trends. The technique was successfully tested at two sites located in Siberia. Forest decline occurred after consecutive droughts during the last decades. Mortality began at hilltops and steep south-facing slopes, shifting with time to lower elevations. Maximum of the desiccated forest area was within steep (18 degrees-25 degrees) south-facing slopes.