Citation

Shu H, Hürlimann M, Molowny-Horas R, González M, Pinyol J, Abancó C, Ma J. Sci. Total Environ. 2019; 693: e133557.

Affiliation

Key Laboratory of Western China's Environmental Systems (Ministry of Education), College of Earth and Environmental Sciences, Lanzhou University, 222 South Tianshui Road, Lanzhou 730000, China.

Copyright

(Copyright © 2019, Elsevier Publishing)

DOI

10.1016/j.scitotenv.2019.07.363

PMID

31369891

Abstract

The effects of land use and land cover (LULC) dynamics on landslide susceptibility are not fully understood. This study evaluates the influence of LULC on landslide susceptibility and assesses the historic and future LULC changes in a high mountain region. A detailed inventory map showing the distribution of landslides was prepared based on the 2013 episode in Val d'Aran, Pyrenees (Spain). This inventory showed that LULC clearly affected landslide susceptibility. Both the number of landslides and the landslide density triggered in grassland and meadow was highest (52% and 2.0 landslides/km²). In contrast, the landslide density in areas covered by forest and shrubs was much lower (15% and 0.4 landslides/km², and 23% and 1.7 landslides/km², respectively). Historical changes of LULC between 1946 and 2013 were determined by comparing aerial photographs. The results indicated that the forest and shrub areas increased by 68 and 65%, respectively; whereas grassland and scree areas decreased by 33 and 52%. Urban area also increased by 532%, especially between 1990 and 2001. Future LULC was predicted until 2097 using TerrSet software. The results showed that the forest area and urban area increased by 57 and 43%, severally; while shrubs, grassland and scree area decreased by 28, 46 and 78%, respectively. Heuristic and deterministic models were applied to create susceptibility maps, which classified the study area into four susceptibility degrees from very low to high. The maps were validated by the 2013 landslide dataset and showed satisfactory results using receiver operating characteristics curves and density graph method. Then, susceptibility maps until 2097 were calculated by the heuristic model and results revealed that landslide susceptibility will decrease by 48% for high-susceptible areas. In contrast, the areas of very-low susceptibility degree will increase 95%, while medium and low-susceptible areas will be more or less constant. This study only includes the effect of future LULC changes on the landslide susceptibility and does not analyze the future impacts of climate changes and the variation of rainfall conditions. Nevertheless, the results may be used as support for land management guidelines to reduce the risk of slope instabilities.

Copyright © 2019 Elsevier B.V. All rights reserved.

Language: en

Keywords

Cover type; Land use; Landslide susceptibility; Modelling; Pyrenees; Slopes