The effect of urban landscape pattern on urban waterlogging

Wu Jiansheng,Zhang Puhua
DOI: https://doi.org/10.11821/dlxb201703007
2017-01-01
Abstract:In recent years, frequent urban waterlogging disasters have aroused great concern. Previous studies have mainly focused on the relationship between land use and waterlogging, but there is not enough information offered in these studies to explore the effects of urban landscape patterns on waterlogging. Taking Shenzhen as an example, this paper used correlation analysis and multiple stepwise regression analysis to explore the effects of landscape pattern on urban waterlogging. As urban waterlogging disaster is a systemic problem, which is related to the water circulation within the watershed, we divided Shenzhen into 56 sub-watersheds as the study units. Then we selected data of 278 waterlogging points during the rainstorm on May 11, 2014 and calculated the waterlogging point density of each sub-watershed to characterize the waterlogging degree. This paper considered 5 landscape pattern indexes to reflect the characteristics of landscape pattern, including LPI (Largest Patch Index), COHESION (Patch Cohesion Index), DIVISION (Landscape Division Index) at classlevel, and CONTAG (Contagion Index), SHDI (Shannon's Diversity Index) at landscape-level. Furthermore, other data, including land use, impervious surface percentage, fractional vegetation coverage, precipitation, topography and storm drainage considered as factors influencing urban waterlogging, were also obtained. The results showed that: (1) Among land use types, the construction land, especially residential land, has the greatest impact on the urban waterlogging disaster, whose growth can significantly increase the disaster degree; (2) At classlevel metrics, the higher level of construction patches' dominance and aggregation and the lower level of landscape division, would lead to more severe waterlogging disasters in a certain area, while the impacts of landscape pattern of greenspace are in reverse; (3) At landscapelevel, the regions with more complex landscape are not prone to waterlogging disaster; (4) Impervious surface percentage and other human factors have greater impact on urban waterlogging than rainfall and other natural factors. This study demonstrated that landscape pattern had significant effect on waterlogging and could provide reference for control of waterlogging and planning management of landscape pattern. © 2017, Science Press. All right reserved.
What problem does this paper attempt to address?