Research On The Landscape Spatial Pattern In Shanghai Waterside Forest Park

According to landscape design principles and results of investigation, the total landscape in Shanghai Waterside Forest Park is divided into 11 landscape zones and 270 patches in the park are classified into 9 patch types. This paper aims to analyze the landscape spatial pattern features and Landscape corridor features by means of FRAGSTATS to calculate some landscape indexes at both the patch level,the class level and the landscape level. Moreover, we discuss the correlation between landscape indexes and scaling effects of landscape spatial pattern. The research result in Shanghai Waterside Forest Park will represent spatial pattern features of small scale artificial landscape which emulates nature. The primary analysis conclusions are as follows:The whole park is distributed and combined by 11 landscape zones that have obviously different features in landscape appearance as well as structure. The whole landscape has high heterogeneity degree and objectively emulates natural landscape. Moreover, the landscape preferably integrates eco-landscape and vision-landscape. The patches in the park have high quantity, the grain of patches has great difference and the shape of most patches is complex. The conjunction between patches is fine, but the patches of the same type dispersed relatively. The original vegetation patches that are reserved in the park have large grain and simple shape. These features express obvious production function. Landscape richness as well as landscape diversity is different according to every landscape zone. Evergreen deciduous broad-leaved mixed forest and evergreen broad-leaved forest are the main patch types. Moreover, evergreen deciduous broad-leaved mixed forest is landscape matrix in the park. Architecture patches and water patches have important service function and ecological function while the other patch types enrich the landscape diversity and raise the view function. In 11 landscape zones, Entrance zone, Rose zone, Erica zone, Orange forest zone, Original forest zone and Waterside zone are controlling-landscape zones, among which exists absolutely advantaged single patch type that controls the landscape appearance and features. The rest landscape zones have 2 or more main patch types, whose landscape appearance is plentiful and landscape pattern is balanced in these landscape zones.The road corridors, hedge corridors and river corridors mutually intervein and connect to constitute corridor-network, which evenly distributes and plays a key role in the formation of landscape spatial pattern in the park. The corridor-network has proper area and density, reasonable distribution and plentiful types, which have positive effect on forming good landscape spatial pattern, exerting landscape function, organizing transportation and appreciating landscape.The correlation coefficients among 17 landscapes indexes that are classified into 5 types are calculated by SPSS. If the correlation coefficient is significant at p=0.01, the indexes have significant correlation among themselves. If the correlation coefficient is significant at p=0.05, the indexes have correlation among themselves. Otherwise, the correlation between indexes is worse. The result shows that in the same type indexes, the correlation coefficients of Shape indexes and Diversity indexes are more significant than Area and Density indexes, Proximity indexes and Contagion indexes have lest significant. Between types of landscape indexes, Diversity indexes most correlate with other types. However, Proximity indexes don't show any correlation with other types and the rest type indexes have fractional correlation among themselves. The correlation coefficients between landscape indexes are influenced by the landscape pattern features, human factors, ecological meanings, mathematical parameters and mathematical units.In constant spatial extent(930×1852m~2),17 landscapes indexes are calculated at different grain sizes in 0.2m,0.6m,1m,2m,3m,5m,10m in every landscape zone and whole park and scaling effects are analyzed by landscape indexes changing which comes from grain size increasing. The result shows that different indexes have different responding with increasing grain size. The changing of the scaling effect curves is unsteady and inscrutable in Area and Density indexes and Shape indexes. The grain effects have definite and clear scaling relations in Proximity indexes and Contagion indexes. The changing of Diversity indexes has less sensitivity. The scaling effect curves are more similar when correlations are stronger among the indexes. The most indexes have critical phenomenon with increasing in grain size. The analysis result of the landscape pattern reflects certain grain size features and the analysis result is valid only under certain grain size.