Research On Spatio-temporal Evolution Of Urban Greenspace Patterns In Typical Pan-pacific Metropolises

Urbanization is one of the main driving factors for vegetation loss and fragmentation in urban area.Thus,as the city prospers,urban planning often emphasizes greenspace management to support urban ecosystems and human health.The Pan-Pacific cities,with nearly 55% of the global urban population,are the fastest urbanizing cores in the world.The urban greenspace is particularly important for this region.However,the lack of consistent long-term land cover information needed for cross-site comparison limits our understanding of greenspace trends within Pan-Pacific cities.In order to detect the spatio-temporal dynamics of the urban greenspace in Pan-Pacific region,we selected 16 major cities across Pan-Pacific region over 27 years,from 1984 to 2012,to explore the spatio-temporal dynamics in greenspace patterns and urban-to-rural gradient trends.We proposed three main questions: In Pan-Pacific region,1)What are the trends in greenspace heterogeneity over time? 2)How does the greenspace heterogeneity vary along urban-rural gradients? 3)How to optimize the greenspace network in this region? In this paper,we collected Landsat composites from 1984 to 2012 and applied the Normalized Difference Vegetation Index(NDVI)thresholding technique to reclassify greenspace into four categories: dense vegetation,medium-dense vegetation,sparse vegetation,and non-vegetated areas.In order to explore the characteristics of the greenspace patterns,landscape indices were calculated with FRAGSTATS.Then,we assessed temporal dynamics of greenspace using the Mann-Kendall test.We also calculated Normalized Distance Index(NDI)to measure urban-rural gradients.Generalized Additive Mixed Model(GAMM)and “Loess” smooth technology were applied to estimate the trajectories of changes in greenspace patterns.We also used Dynamic Time Warping(DTW)to measure the differences of those trajectories.Finally,Maximum Information-based Nonparametric Exploration method(MINE)was used to detect the paired correlation between the changes in greenspace patterns and social development factors,temperature,precipitation and greenspace policies.Our results show,from 1984 to 2012:(1)The Normalized Difference Vegetation Index(NDVI)in Pan-Pacific cities increased slightly over the past 28 years.Specifically,dense vegetation(most of which are urban forest or natural vegetation)and sparse vegetation(anthropogenic greenspace)increased and their patterns became more connected,contrasting the medium-dense vegetation(most are seminatural greenspace).Pan-Pacific cities have diverged into “greening” and “browning” groups according to the increasing or decreasing trends of greenspace.(2)Dense vegetation showed an increasing trend in both greening and browning cities,accompanied by more connected patches and more complex patch boundaries.Contrary to greening cities,the proportion of medium-dense vegetation in browning cities declined and the patterns became more fragmented.Furthermore,dense vegetation increased from the city centre to suburban area in greening cities,while there was no obvious urban-to-rural gradient trend in browning cities.The greatest long-term changes in greenspace were evident in the outskirts in both greening and browning cities.(3)In contrast to North American cities,dense vegetation in Asian cities increased and the patches became less fragmented,and the patterns of medium-dense vegetation,sparse vegetation and non-vegetated areas became less connected.The urban greenspace in both Asian and North American cities had similar urban-to-rural trends.Specifically,dense vegetation and medium-dense vegetation increased from the city centre to city fringes while sparse vegetation increased.(4)For Chinese cities,the dense vegetation in Changsha,Dalian and Shenzhen and medium-dense vegetation in Nanchang increased and became more connected.In contrast to coastal cities(Dalian and Shenzhen),sparse vegetation in Nanchang and Changsha increased and became less fragmented.The urban-to-rural gradients of greenspace in selected Chinese cities had three different trends: the first one was a parabolic trajectory(“∩or∪”),second was a horizontal line(“—”)and third was monotonic increasing or decreasing.Income,gross domestic product per person and urban expansion showed a non-linear relationship between the changes in greenspace.(5)For the future construction of greenspace network in Pan-Pacific region,changes of greenspace in suburban areas and city fringes should not be ignored: we should pay more attention to the construction(especially for browning cities)and conservation(especially for greening cities)of dense vegetation in these regions.Furthermore,we should improve the proportion of medium-dense vegetation in the suburban area in greening cities and the city centre radius in browning cities.For sparse vegetation,all greening and browning cities should focus on the construction of corridors and strengthen the connectivity between patches.The consistent or different changes in greenspace heterogeneity between greening and browning cities,Asian and North American cities provide an important baseline and new spatial data for decision-makers to set up regional strategies and planning for green infrastructure.