Spatial pattern variation of artificial sand-binding vegetation based on UAV imagery and its influencing factors in an oasis–desert transitional zone

Abstract

The spatial pattern of vegetation can reflect the impacts of the environment on plants and the response of plants to the environment, which can promote a deep understanding of the potential driving mechanisms of vegetation evolution and community maintenance. A sand-binding vegetation system without irrigation has been implemented in the oasis–desert transitional zone since the 1970 s, where the annual precipitation has been approximately 120 mm. While the mobile dunes have been effectively stabilized, a patchy pattern of sand-binding vegetation has been observed. However, we do not understand why the pattern of sand-binding vegetation changed from the initial uniform distribution to the current patchy pattern. In this study, low-altitude UAV remote sensing technology and imaging-based quantification techniques were used to explore the effects of biotic and abiotic factors on the spatial patterns of sand-binding vegetation over 50 years. The spatial pattern of Haloxylon ammodendron changed gradually from a uniform distribution to an aggregated distribution, and the degree of patch fragmentation of H. ammodendron at the landscape scale gradually increased with the age of the sand-binding vegetation. The artificial sand-binding vegetation composed of H. ammodendron showed discontinuous change in which the system state reached a transition point after 30 years and changed to another state after 40 years. There were no significant correlations between the landscape indices and soil water content in the shallow layers (0–10 cm, 10–50 cm), while the soil water content in the 50–100 cm layers was significantly negatively correlated with the class area, percentage of landscape, largest patch index, percentage of like adjacencies and aggregation index and was positively correlated with the normalized landscape shape index. The soil water content in the 100–200 cm layers was positively correlated with the number of patches and patch density. Competition intensity at the individual level had a more significant effect on the area-type indices, and competition intensity at the population level had a more significant effect on the clustering-type indices. Finally, we found that the soil water content in deep layers and competition are the main drivers of the H. ammodendron spatial pattern change from a uniform pattern to a patchy pattern. These findings enrich theory on the self-organization of vegetation in arid and semiarid environments and have important theoretical and practical significance for the establishment and management of artificial vegetation in arid areas.