How do physical and social factors affect urban landscape patterns in intermountain basins in Southwest China

2021 
Urbanization in mountainous areas concentrates in intermountain basins (locally named as “bazi”) formed by various geomorphological processes, resulted in a unique urban landscape patterns (ULP) and hierarchically structured determinants. Detecting ULP in basins across the mountainous Yunnan Province in Southwest China; disentangling the impacts of physical and social drivers on ULP in basins of variable sizes and administrative levels. Five indices of ULP were extracted from Landsat 8 OLA, and linked to causal variables of basin geomorphology, land cover types, administrative structure, road network, within-basin topography and climate; and generalized linear regression and random forest models were applied to estimate factor importance for variation of ULP within basins. (1) The capital cities of Yunnan Province and its 100% prefectures, 70% counties and 30% townships were located inside a total of 1198 basins with the area ranging from 1.0 to 1385.1 km2. (2) ULP patterns in basins of county and higher administrative levels were prominently constrained by physical and social features of basins, and were less interpretable in basins of lower administrative level. (3) Basin area but not shape showed dominant impact on ULP, with direct limit effect revealed only in basins of small size and high administrative level. The administrative structure dominated the spatial configuration of ULP (i.e. patch density, connectance and agglomeration) in basins, and regulated the role of road system, topography and climate with hierarchical differences. The ULP in mountain areas was co-dominated by physical and social features of basins. Mountain urbanization impacts should be reduced by rational urban development planning harmonizing physical constraints and administrative optimization.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    64
    References
    3
    Citations
    NaN
    KQI
    []