Article ID Journal Published Year Pages File Type
6303367 Journal of Arid Environments 2016 8 Pages PDF
Abstract

•11 indices can represent 43 landscape metrics through PCA method.•The key factors are identified through detrended canonical correspondence analysis.•The ecological factors are the key factors in determining landscape metrics.•There are obviously spatial scale effects exist in analysis of landscape metrics.

Landscape patterns in agro-pastoral ecotones are influenced by natural factors and human activities. However, the main factors that influence landscape metrics of agro-pastoral ecotones have not been fully elucidated. To further understand conditions influencing landscape formation, we conducted a series analysis to explore the relationship between ecological factors (annual average precipitation, annual average temperature, NDVI, altitude, aspect, slope, curvature, land use, and human disturbance) and landscape metrics pattern in the range of 100-5000 m spatial extent, within an agro-pastoral ecotone in Inner Mongolia, China. Using principal component analysis and the detrended canonical correspondence analysis from 43 landscape metrics, we successfully identified several key factors that determine the landscape metrics values. Agriculture and livestock grazing land use induce the landscape configuration to become homogeneous and simple. Nevertheless, our results show that the influences of human activities on landscape metrics are surprisingly not strong. Rather the natural ecological factors, in particular temperature, precipitation and altitude, had the greatest influence on landscape metrics values. This study provides a theoretical case for the scaling effects and develops techniques for identifying the key ecological factors influencing on landscape metrics, so as to improve landscape management decisions in semi-arid regions and other ecotones.

Related Topics
Physical Sciences and Engineering Earth and Planetary Sciences Earth-Surface Processes
Authors
, , , ,