CollectHomepage AdvertisementContact usMessage

Arid Land Geography ›› 2022, Vol. 45 ›› Issue (3): 946-954.doi: 10.12118/j.issn.1000-6060.2021.328

• Regional Development • Previous Articles     Next Articles

Morphological types division and its formation of dispersed rural settlements based on Alpha Shape: A case of Long Town, Mizhi County

WANG Tianyu(),HUI Yi’an(),RUI Panpan,SHI Ying   

  1. College of Urban and Enviromental Sciences, Northwest University, Xi’an 710127, Shaanxi, China
  • Received:2021-07-19 Revised:2021-12-10 Online:2022-05-25 Published:2022-05-31
  • Contact: Yi’an HUI E-mail:1017743096@qq.com;596060353@qq.com

Abstract:

To quantitatively classify the morphological types of dispersed rural settlements and explore the reasons for their formation, this study cites Long Town, Mizhi County, loess hilly-gully region in northern Shaanxi of China as an example, attempting to use Alpha Shape algorithm to extract the settlement contour, calculate and classify the settlement contour type using the Boyce-Clark index, and finally explore the factors that affect the settlement contour by combining numerous linear regression and regression tree methods. The results show that: with a rolling radius of 100 m, the Alpha Shape algorithm can extract 76.1% of the accurate contours of the settlements; the types of rural settlements in the study area can be divided into five types: block rectangle and its variants, extension, dumbbell and bead, ribbon rectangle and its combination, and linear. The Boyce-Clark index increases continuously; elevation and shape index are positively connected, water availability and shape index are negatively connected in high altitude areas, settlement area and shape index are first negatively connected and then positively connected. This study provides a reference for the delineation of the living space boundary of dispersed settlements and enhances the understanding of dispersed settlements in loess hilly-gully areas.

Key words: rural settlement, morphological type, Alpha Shape algorithm, Boyce-Clark index, loess hilly-gully region