Abstract
Abstract. Urban area hotspots can be considered as an ideal representation of spatial heterogeneity of human activities within a city, which is susceptible to regional urban expansion pattern pattern. However, in previous studies most researchers focused on extracting urban extent, leaving the interior variation of nighttime radiance intensity poorly explored. With the help of multi-source data sets such as DMSP/OLS (NTL), LST and NDVI, we proposed an applicable framework to identify and monitor the spatiotemporal trajectory of polycentric urban area hotspots. Firstly, the original NTL dataset were calibrated to reduce inconsistency and discontinuity. And we integrated NTL, LST as well as NDVI and established an urban index TVANUI capturing the approximate urban extents. Secondly, multi-resolution segmentation algorithm, neighborhood statistics analysis and a local-optimized threshold method were employed to get more precise urban extent with an overall accuracy above 85% and a Kappa above 0.70. Thirdly, the urban extents were utilized as masks to get corresponding radiance intensity from calibrated NTL. Finally, we established the Gaussian volume model for each cluster and the resulting parameters were used to quantitatively depict hotspot features (i.e., intensity, morphology and centroid dynamics). All the identified urban hotspot showed our framework could successfully capture polycentric urban hotspots, whose fitting coefficients were over 0.7. The spatiotemporal trajectory of hotspot powerfully revealed the impact of the regional urban growth pattern and planning strategies on human activities in the city of Wuhan. This study provides important insights for further studies on the relationship between the regional urbanization and human activities.
Highlights
Urbanization is a complicated phenomenon involving massive population shift, spatially expanded built-up area and adjustment of industrial structure (Cohen, 2006)
Radiation and atmospheric correction were applied to Landsat ETM+ imagery and the maximum likelihood method classification (MLC) was used to obtain land use type of Wuhan, the built-up derived from MLC was regarded as a reference for precision validation
Considering the study about dynamics of Wuhan urban hotspots was started from 2000, the Xinzhou district was discussed in subsequent analysis
Summary
Urbanization is a complicated phenomenon involving massive population shift, spatially expanded built-up area and adjustment of industrial structure (Cohen, 2006). DMSP NTL imagery has been extensively used as a powerful tool for mapping urban extent despite its weaknesses of blooming and saturation (Croft, 1973; Elvidge et al, 1997; Yu et al, 2015; Xie et al, 2014; Cao et al, 2009). As main driven force of landscape change, has greatly influenced land-use type which results in higher intensity emerging in urban areas namely nighttime urban hotspot as an ideal proxy for describing spatial heterogeneity of human activities (Zheng et al, 2018). In order to make better use of multi-source data to identify Wuhan's poly-centric urban structure, LST, NDVI and NTL were integrated to establish an urban index termed the Temperature and Vegetation Adjusted NTL Urban Index (TVANUI) which is closely related to urban features and mitigates the NTL blooming and saturation for mapping urban area (Zhang X et al, 2018). Gaussian volume model was used to explore urban structure of Wuhan
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
