Fig 1.
Location map of Nansha district, Guangdong province, China.
Nansha district consists of two sub-districts and six towns. The locations of urban and rural communities are based on the Comprehensive Planning of Nansha New City (2012) [35]. The map was generated using the free and open source software QGIS version 2.18 (http://www.qgis.org/en/site/about/index.html).
Table 1.
Information on 24 selected typhoon samples.
Fig 2.
Sample results of typhoon Utor.
(a) maximum grid wind distribution (unit: m/s); (b) maximum grid hourly precipitation (unit: mm/h); resolutions of both are 1 × 1 km. Wind speed modeled in this study is at 850 hpa height because wind speed at 850 hpa height is considered surface wind in meteorology and has the greatest effect on surface features, such as buildings and infrastructure. The map was generated using the free and open source software NCAR Command Language version 6.4.0 (2017) (http://dx.doi.org/10.5065/D6WD3XH5).
Fig 3.
Modeling results for Utor in Nansha.
(a) distribution of maximum wind speed (unit: m/s); (b) maximum precipitation (unit: mm/h) in each grid square in Nansha; both have resolution of 1 × 1 km. The map was generated using the free and open source software QGIS version 2.18 (http://www.qgis.org/en/site/about/index.html).
Fig 4.
Threshold value and parameter estimation of typhoon maximum wind speed in sample grid.
(a) Mean residual life plots; (b) Re-parameterized scale parameter; (c) Shape parameter. Approximate straight line in (a) from 9 to 12 implies that suitable threshold value should be around 9, and similar trends of two parameters are also presented in (b) and (c).
Table 2.
Selected indicators for resilience assessment.
Fig 5.
Hierarchical structure of composite resilience index in Nansha district.
Table 3.
Random index (RI) values.
Fig 6.
Results of extreme value analysis.
(a) wind speed distribution (unit: m/s); (b) precipitation map (unit: mm/h); both are at 100-year return level. The map was generated using the free and open source software QGIS version 2.18 (http://www.qgis.org/en/site/about/index.html).
Fig 7.
Diagnostics from the fitted generalized Pareto distribution function (modeled dataset) to the WRF-simulated maximum wind speeds (empirical dataset).
(a) quantile plot, (b) density plot, and (c) return level plot.
Table 4.
Resilience components in first level, extracted factors in second level, and primary variables of each factor.
Table 5.
Pairwise comparison matrix: Extracted factors in second level.
Table 6.
Pairwise comparison matrix: Four components of resilience in first level.
Table 7.
Relative weights of factors in second level and components in first level.
Fig 8.
Composite index of resilience to typhoon disaster in Nansha district.
a) sub-district level; b) 1×1-km grid level (empty grids represent grids with zero population). The map was generated using the free and open source software QGIS version 2.18 (http://www.qgis.org/en/site/about/index.html).
Fig 9.
Percentage of grids in sub-districts with z-scores in different resilience classes and average class of each sub-district.
Fig 10.
Maps of resilience component score.
a) social component; b) economic component; c) infrastructural component; d) natural component; classified as low (< −1.0 SD), medium-low (−1.0 to 0.0 SD), medium-high (0.0 to 1.0 SD), and high (>1.0 SD). The map was generated using the free and open source software QGIS version 2.18 (http://www.qgis.org/en/site/about/index.html).