The Spatiotemporal Evolution And Spatial Matching Of Urban Green Space In Hangzhou Bay

With the sharply rising of living standards, city residents have proposed a higher need for the improvement of urban environment. Studies about the spatiotemporal variation of urban green space can expand and deepen our understanding of the evolution of urban environment under rapidly urbanization, and eventually provide meaningful guidance for the optimization of the distribution of urban green space in order to promote a full use of its ecological, social and economic benefits. Urban green space is widely distributed, fragmentized, and highly dynamic because of intense human disturbances. Therefore, accurate extracting of urban green space in a complex urban environment requires comprehensive consideration of the temporal and spatial resolution of remote sensing data and to adopt appropriate image analysis techniques. Using Landsat images as main data, our study applied sub-pixel urban green space information extraction and change detection techniques to investigate the spatiotemporal dynamic and landscape pattern of urban green space in six main cities in Hangzhou Bay, a typical rapid urbanizing area in China, from 1990 to 2013. Furthermore, with the guidance of environmental justice theory, we analyzed the status and variation of spatial disparity between urban green space and urban population and related environmental justice issues at street-town level. The major contents and conclusions were summarized as follows:(1) Using regional green space extraction technique based on Multiple Endmember Spectral Mixture Analysis, urban green space in main cities of Hangzhou Bay in 1990,2002 and 2013 were extracted at sub-pixel level. By applying the pure vegetation mask based on threshold method first, the model computation and the disturbances in pure vegetation cover area caused by subtle difference in vegetation spectra were effectively reduced. Adding the water endmember and shade endmember to the standard VIS model can improve the performance of the model in cities with typical characteristics of canal towns south of the Yangtze was significantly improved. The green space around the wetland, farmland water nets and water inside the cities can be accurately extracted. When choosing the spectra for different complexity level of models, only the spectra with good performance in two endmember models were used to run three endmember and four endmember models, which can greatly reduce the computational complexity while ensure accuracy. Regional green space extraction technique based on Multiple Endmember Spectral Mixture Analysis could acquire accurate information of urban green space, especially for those small and fragmented patches within the urban area and linear green space on the sides of the streets and rivers. These information provided a steady foundation for our understanding of the urban environment.(2) RGB fraction models were developed based on the vegetation fraction maps of 1990,2002 and 2013. Based on RGB additive principle, minimal changes and spatiotemporal changing hotspots of urban green space were visually expressed through the rich color combination. Over the past 20 years, the green space in the main city region of the Hangzhou Bay has experienced a drastic change. Loss of green space mainly occurred on the edge of the built up area, with a severer degree and range between 2002 and 2013. By contrast, green space in the urban center and coastal areas showed a recovery.The hierarchical fraction based landscape index was utilized to analyze the changing landscape pattern in the city region, old city district and the urban-rural gradient from quantity, fragmentation, dominance, concentration, and diversity. At the city region level, the green space landscape is gradually degrading from full vegetation cover landscape into no, low, and middle vegetation cover landscape. Intensive human interference, especially the severe spread and expansion of urban build up area, cut off the connection between natural landscapes, resulting in a fragmented, scattered, and disordered regional green space landscape. Differently, a recovery of the green space landscape in the old city district of the six cities emerged, especially from 2002 to 2013. The increase of small and scattered green space patches resulted in a significant increase of middle and low vegetation cover green space and decrease of no vegetation cover green space. The formation of urban green space network splitted the large impervious patches into smaller ones, resulting in an increase of green space landscape diversity. The extinct pattern of green space landscape variation between two periods revealed the impact on the urban environment of different green policies. The new urban green strategy and innovative greening measures promoted the growth of the amount and spatial distribution of middle and low vegetation cover green space, which played a positive effect on the improvement of the urban environment. Gradient analysis further revealed the spatiotemporal pattern of urban green space along with the expansion of build ups.(3) Based on the census data, Gini coefficient and Theil’s index were used to analyze the spatial disparity between the distribution of green space and public parks and the distribution of urban population at street-town level from 2000 to 2010. Along with the urbanization process, the disparity between green space and urban population increased while between public parks and urban population decreased. In larger cities, the disparity in green space exceeded the public parks and became a more prominent regional environmental issue. By comparing the urban group and rural-town group, a more prominent disparity of green space was identified in the urban group, as well as a non-negligible urban-rural differences. While the disparity of public parks mainly lay within rural town group. Due to the progress in the construction of public parks in recent years, especially the community parks and village parks, the disparity of public parks within rural-town group and the urban-rural differences has been considerably relieved.Aging people and floating population showed difference in the variation of the disparities. The disparities of green space and public parks within aging population are larger than the total population, while the floating population had smaller disparities. The specific characteristic of spatial aggregation of these two population groups might explain these differences. By further calculating the per capita green space and per capita public parks in the urban group and rural-town group, a significant urban-rural difference in acquiring green space and public parks existed among urban residents, especially for aging population. And the floating population as a whole was at a disadvantage in acquiring public parks by compared with local residents, despite the smaller difference within their group. The future urban green space planning should adhere to the people-oriented principle and give thorough consideration to the needs of different groups in order to finally realize efficient allocation of environmental resources and fair distribution of social welfare.Rapid urbanization has a profound and complex influence on the regional urban green space landscape. A flourishing economy activities and the rapid growth of population stimulated the demand of the urban land, posing a great pressure on green space inside and surrounding the city. At the same time, the growth of awareness of environmental protection and progress of greening concept may improve the layout and management of urban green space. Studies on the spatiotemporal response mechanism of urban green space for the rapid urbanization at multi-scale will improve our understanding of the status and evolution of the urban environment. Information extraction and landscape analysis method based on sub-pixel level can obtain information about the spatiotemporal variation and landscape pattern changes which were difficult to be extracted through traditional analysis methods, providing a new train of thought for application of moderate resolution remote sensing images in the urban landscape ecological studies.