基于网络街景的城市森林测量、评价及城市间分异规律研究

Urban environment deterioration in China requests urban vegetation (mainly composed by urban forest and tree) exerting larger ecological service, while one basis for policy makers is an overall, exact and efficient evaluation of urban forests both inter-cities level and different districts within a city. At present, field survey as the main method is a time- and labor-consuming work, whether or not internet street-view(ISV) data are possible good enough for such information derivation is the main focus of this proposal. Thus, through comparison of different “moving ruler” references, different gathering angles of street-view pictures from different companies, cross-checks of different image processing software and re-programming development, we want to find the method and regression equations for measuring DBH(diameter at breast height), tree height, canopy size, and forest density, as well as visual inspection possibility for detecting tree species composition, forest vertical structure, tree health status and management risks. As the second phase of this study, at least 4 cities will be selected to improve the general applicability of this ISV-based method. We want to know which part of parameters from this method are most reliable throughout different cities, and how to improve the data derivation efficiency. We also want to know the method of carbon sequestration estimation and green quantity estimation from this ISV-based method. As the last part of this study, ISV-based method will be used to make a whole-China city comparison (at least 10 cities). By measuring the most reliable parameters from ISV, we want to know the intercity and intra-city differences in forest composition, structure of arbor-shrub-grass layers,carbon sequestration, green quantity, forest health and management risks. Through a comprehensive analysis between this data and latitude-related natural differences, urban-rural gradient and social-economic parameters, we want to find the underlying rules for this differences. Hopefully, this study can provide a much rapid and economic method for urban road-side forest evaluation, and the underlying rules clarification will favor urban forestry and greening practice and management.

我国恶化的城市环境要求以树木为主体的城市植被发挥更大生态服务功能，城市间及市内不同区域间全面、准确、高效的评价比较是管理决策的基础。现行实地调查方法费时费力，能否通过网络街景数据实现森林测量与评价是本研究的核心。基于此，对比不同移动标尺参照物、网络街景照片采集方法及来源、不同图片处理软件及再开发对测量的影响，确定测量树木胸径、高度、冠幅、密度的经验公式以及视觉评定林分组成结构、管理风险的可靠性和准确度；进一步通过多个典型城市实测，验证提升技术方法，明确最可靠普适性测定指标，形成可行性高、提升效率的三维绿量、绿色指数及碳汇功能估计的标准方法；最后，基于上述方法评价全国10个以上城市森林树种组成与结构、碳汇和绿量功能以及健康和管理风险，揭示我国城市森林变化规律及与纬度、城市化和社会经济的耦合机制。预期结果能够为城市森林普查提供更简单实用的方法，为我国城市造林绿化与管理对策提供理论与技术支撑。

森林树木大小测量是其功能评价的基础，开发基于网络大数据的测量方法将能够提高测量效率、提升评价城市森林功能的准确性。本项目超额完成了相关规定认为，原定11个城市测量、实际完成最多35个城市。文章主要成果产出22篇文章，是原计划的4.5倍；并获得省级一等奖1项、出版科学出版社专著1部。建立了一个基于街景数据对城市森林测量方法，明确了其精度和可靠性，即：胸径、树高(树高和枝高)和冠层大小的精度分别为92%、87%和80%，不同人测定结果重复性很高 (r2 > 0.92)，不同人测量的最大差异：胸径0.6 - -1.9厘米(平均在22厘米)、枝下高8-50厘米高度(平均值在8米)；这种方法所需劳动力成本和时间是实测的三十分之一到十分之一。树木个体大小完成了11个城市的测量，明确了纬度变化规律：及从南到北纬度增加、树木个体小的树木所占比例增大、树木密度也增大。制定基于街景数据对城市森林结构组成（垂直结构、树种组成、大小组成）、生态功能评价和城市森林健康及风险（叶色正常与否、支架与否、枯梢与否）评价方法体系；用于全国35个城市多个城市的评价。明确了城市化影响规律（树木个体大小），提出未来森林生态服务功能评价需要考虑“城市化导致大树增多、密度降低，而土壤碳累积增多”，构建适应性强、弹性城市的绿地系统也需要考虑城市化引起的林分自我更新能力变弱问题。建立了32个城市的城市街景绿色程度数据库，发现其经纬度变化规律及社会发展影响：即经纬度等地理气候决定最大绿色程度，而社会发展决定其公平性。原计划培养 3 名研究生。现已经培养完成博士后1名（张丹），毕业博士2名（吕海亮、王琼），毕业硕士8名（萧路、周伟、张波、王洪元、张菁华、刘晓、王可、杨艳波）；在读博士生2名（萧路、田盼立）。是原计划的4.3倍。课题人才培养显著，新进中科院青促会成员1名，副研究员3名。相关结果为城市森林的快速测量和评价奠定了基础。