OBJECT-BASED CLASSIFICATION OF LANDSCAPE INTO LAND MANAGEMENT UNITS (LMUs)

A cost effective, reproducible and reliable system of classification of landscapes into homogeneous units characterizing land based on its physical, biological and cultural characteristics -denoted as Land Management Units (LMUs) is of significant importance for management and planning purposes. Especially for developing countries, LMU systems are very critical as they aid in decision making process for achieving sustainable development and conservation of resources. Management and planning act requires understanding of common and distinctive properties of sites, capability of land to produce resources and respond to management decisions, hence it is essential to reduce enormous amount of environmental variability into conceptual units that reveal capacity and capability of land. This study aims to develop a method of Geographic Object-Based Image Analysis (GEOBIA) to generate LMUs at a biodiverse region dominated by forests, shrublands and endemic species which is subject to degradation and requires natural conservancy in Blacksea region in Turkey. Starting point for classifying landscape into LMUs is to explore landscape as groups of biotic and abiotic factors representing ecological units that have discernible structure, function, geo-morphology and disturbance regimes. Topography as a controlling mechanism and almost stable structure plays predominant role in determining ecological conditions and it addresses the question of how to best define landform based landscape entities. The proposed system employs object-based analysis of Digital Terrain Models (DTMs) to derive landform elements and utilizes landform-landcover assemblages to perform ecological regioning. Those conceptual objects that form a framework of connected entities are then enriched given other available information e.g. land use, ownership and hazards to ultimately end up with map units called LMUs.