Study On The Spatial Distribution,Morphology And Human Pressure Of Ponds In Urban Built-Up Areas

Wetlands have a high ecological value and are extremely important to the ecological environment and biodiversity.As the most common and widely distributed wetland type in urban areas,small urban water bodies provide important support for building a better urban ecological environment.A pond is a small,shallow,static or slow-flowing body of water with a variety of shapes and contours,formed either naturally or artificially.It is a special type of wetland and an important part of small urban water bodies with rich ecological and environmental benefits and biodiversity functions.However,at this stage,there are still many problems in the management and protection of urban pond resources,the public awareness of the ecological value of ponds still needs to be improved,and relevant pond studies are still limited to discovering the economic value and other aspects,ignoring the ecological service value of the ponds themselves,and various unfavorable factors have led to serious pollution or landfill of reservoirs in urban areas to free up land resources for urban construction.Therefore,it is urgent and valuable to conduct timely research on urban ponds in China’s metropolitan areas in order to protect and manage them as well as to bring into play their positive role for sustainable urban development.The Pearl River Delta（PRD）region faces the well-developed water system of the South China Sea,and the retention of pond resources is high.The formation of Lingnan characteristic water culture in some regions is associated with the existence of ponds,while the rapid urbanization expansion in the PRD region poses a certain threat to the ecological environment of ponds,so it is typical and representative to study the pond resources in this region.Based on this,this paper combines object-oriented and deep learning methods to spatially extract and monitor the ponds within the built-up areas of nine cities in the PRD,to explore the feasibility of spatial information extraction of ponds in complex environments,and to analyze the spatial distribution,morphology,and human pressure with respect to the attributes and characteristics of the pond landscape and the surrounding environment.From the analysis and research,the following main conclusions are drawn in this paper.（1）Combining object-oriented and deep learning methods can extract ponds in urban areas comprehensively and with high accuracy.In the complex subsurface environment of the city,the powerful convolutional neural network and learning ability of the deep learning model make the extraction experiments can still extract the ponds in the urban environment under the influence of various disturbing factors（shadows）,and the overall accuracy is high,and this technical solution effectively alleviates the shadow problem of urban water extraction;（2）The analysis of the spatial distribution of ponds reveals that there are spatial differences in the retention of ponds among the cities in the study area,including large differences in the number of ponds.Combined with the study on the clustering effect of ponds,this paper attempts to construct a "blue rate" index based on small urban water bodies,which is derived from and refers to the urban "green rate",and the "blue rate" index of Foshan and Zhongshan is higher and the distribution of ponds is more concentrated;（3）The analysis of spatial morphology of ponds reveals that the morphological characteristics and types of ponds are correlated,and there is a correlation between spatial morphology and distribution.This paper proposes a quantitative discrimination method for typical pond types,and the discrimination method is feasible.This paper classifies pond resources into 3 levels and 23 types of ponds,with the criteria of formation mechanism,topographic location and specific use.The morphological characteristics of ponds are analyzed by the orthogonal image index and combined with the pond classification study,and it is found that the morphological characteristics of ponds are correlated with the pond types,and a quantitative discrimination method is proposed and experimentally tested for three typical pond types;（4）The analysis of human pressure on the ponds revealed that the maximum values of human pressure on the ponds were generally within the core urban area and lower in the non-core urban area,but some outliers appeared,and the existence of the outliers was related to the corresponding feature types.By introducing the intensity of ecological anthropogenic disturbance and combining the two indicators of shoreline ratio index and nighttime light disturbance,a comprehensive analysis of the human pressure on ponds in the study area was conducted,and a cluster analysis was performed with the help of the results of the comprehensive human pressure on water ponds.The results of the cluster analysis show that low human pressure outliers exist in the core urban areas,generally such as urban waterfront parks or wetland parks,while high outliers also exist in the surrounding non-core urban areas,and these high outliers are generally distributed in areas such as warehouses,industrial industrial parks,and suburban villages on the outskirts of the city.