Representative landscape indices of biological soil crusts distribution in the hilly Loess Plateau of China

Abstract

Landscape indices can quantitatively describe the distribution characteristics of biological soil crusts (biocrusts). However, there are too many landscape indices, with high redundancy. We investigated 58 plots of biocrusts with different distribution patterns in the Hegou watershed of Wuqi County, Shaanxi Province, located in the hilly Loess Plateau. First, we calculated 15 common landscape indices, and selected representative landscape indices that could describe the biocrust landscape pattern and had specific ecological significance, based on correlation analysis, factor analysis, and sensitivity analysis. The reliability and rationality of the representative landscape indices were verified with data of the different biocrusts coverage in the Yingwoshanjian watershed of Yangjing Town, Dingbian County, Shaanxi Province. The results showed that 10 of the 15 landscape indices had significant correlations. Total edge (TE) and edge density (ED) were not significantly correlated with number of patches (NP), patch density (PD), clumpiness (CLUMPY), and interspersion juxtaposition index (IJI), respectively. The percentage of landscape (PLAND), ED, patch cohesion index (COHESION), and splitting index (SPLIT) described the spatial distribution characteristics of biocrust from coverage, length, connectivity, and fragmentation, respectively. The cumulative contribution of the three common factors represented in describing the spatial distribution of biocrusts was 91.6%. The study identified the representative landscape indices that could quantify the complexity of biocrusts distribution and thus would provide a theoretical basis for studying the pattern evolution of biocrusts and their relationship with ecological processes.