Using digital elevation model, satellite and ground data to mapping of vegetation cover diversity
Conventional methods of forest monitoring, such as field observations and thematic mapping, are often insufficient to meet the needs of today’s research tasks for information on vast and highly heterogenic land cover fragments containing forest areas with rapidly changing boundaries. For this reason, a concept of ecosystem monitoring based on space imagery analysis combined with application of mapping methodologies, GIS, and ground data has been developed as an alternative approach over the recent decades.
We have attempted to develop a methodology to classification and mapping of vegetation cover using satellite data, digital elevation model, ground observation data and GIS.
A test site area (68º - 69º N, 88º - 89º E) was located in the Northern Siberia boreal forest.
To identify vegetation type Landsat 8-OLI images were classified by the method of maximum likelihood. In the process of images classification we obtained land cover classes: forests and woodlands, forest-bog complexes, tundra, tundra-forest complexes, bogs, stones.
To correctly classify and understand the position of any one vegetation type in a vegetation development series and to assess its current state, one needs to know a vegetation growing conditions controlling the probability of occurrence of any vegetation community type. The classification of potential vegetation growing conditions was carried out: two-layer DEM-composite (elevation above sea level and slopes) was classified using ISODATA. Land cover classes relatively similar in relief morphometric parameters were identified as geomorphological complex of vegetation growing conditions. To perform more detailed classification, we analyzed each geomorphological complex and identified sites that were homogeneous in slope. Those sites corresponded vegetation growing conditions types. The electronic layer of potential growing conditions types was created as a basis of vegetation cover mapping.
Then the expert system for classification and mapping of vegetation cover was developed using Knowledge Engineer module / ERDAS Imagine. The initial classes obtained from remote sensing data classification were distributed using a preliminary classification of growing conditions types and the associated vegetation. As a result, 16 vegetation types or their complexes were obtained.
The obtained maps reflects the diversity of vegetation cover (level of vegetation types and their complexes) and vegetation growing conditions types. They can predict the pace of regeneration succession in a range of vegetation growing conditions. Therefore, these maps can be used in the complex models of carbon balance and carbon pools estimation.