An algorithm was developed for assessment of long-term variations of vegetation characteristics. The algorithm use NDVI data from spectroradiometer MODIS of Terra and Aqua satellites. The algorithm includes pre-processing procedures – the restoration of missing values, smoothing using Savitsky-Golay filter. To analyze the areas with snow cover the minimum length of the growing season is determined for the full length of the time series. Obtained after removing the seasonal component NDVI values are used to construct a linear regression and determine the trend.

As a result of applying the algorithm the map of the spatial distribution of NDVI linear trends was created from 2000 to 2016 for the West Transbaikalia. Examples of visual verification of vegetation cover changes using satellite images of ultra-high spatial resolution are presented.