Academic Track Paper

Landscape diversity indication from remote sensing data using open source software


Tetyana Kuchma


Presenter Biography: 

I have master degree in ecology and 10-year experiance working with RS and GIS in private companies and scientific-and-research institutions in the field of agriculture, climate change, landscape analysis and modeling, risk assesment and route optimization. Since 2010 I am the course on GIS application in land management and ecology, in particular for students of National University of Kyiv Mohyla Academy (Ukriane). Currently I am working on PhD tesis on Landscape diversity indication using remote sensing.


The paper is focused on landscape diversity indication using remote sensing data in odder to solve the urgent problem of biological and landscape diversity loss. An excessive fragmentation of natural vegetation cover into isolated areas and insufficient density of erosion control facilities and protection network, the dominance of monoculture in agriculture, the high proportion of arable land, and plowing of grasslands and steppe areas to the boundary of the forests or river in the landscape are observered, causing the reduction of environmental sustainability of agroecosystems. The landscape structure assessment tools for selecting the most appropriate models of rational spatial organization of agroecosystem, conserving biodiversity and ecological stability, is required at both the regional administrative land management level, and at the level of farm. Assessment tool must be simple, efficient and quantitative objective, so that it could easily be used as by experts and local authorities so as by farmers. About hundred indexes for landscapes diversity assessment are developed and described in scientific literature. But they often give a different or even an opposide estimates of diversity. Thus there is a need to justify the objective indices of landscape diversity, effective for agricultural landscape monitoring. Of particular importance is also a way of grouping elements of the structure, et classification of remote sensing data. Our approach is based on remote sensing data procesing using BEAM Visat open source software to determine the classes of agro landscape structure (land cover), and subsequent calculation of landscape metrics to select of the most appropriate ones using QGIS and R software. Three test areas, differing considerably in structure (with varying degrees of plowed territory), were selected for the comparative calculations of landscape metrics. The following landscape metrics were demonstrated to be the most effective: habitat diversity and heterogeneity, total edge length, portions of natural, semi-natural and intensive land used, Shannon diversity index, mean patch size index, and inter-dispersion index.

? Top