Context Internet-based street-view greenery (SG) is a new tool for evaluating urban green infrastructure, with proved vital services for residents, while until now, no report is on SG-aimed management from forest structure and landscape patterns. Objectives To find out plant composition, tree size, and landscape pattern's contribution to inter-and intra-city SG variations, and implication for SG-maximization management. Methods The SG was quantified by upper green view index (sky GVI), middle GVI, and ground GVI by using Baidu Street View in Harbin and Changchun; forest structure and landscape metrics were also measured in field survey, street-view picture, and remote sensing, then matched to 2 km x 2 km grids. Sampling density was 0.25-3 plots/km(2), securing uncertainty < 10%. Results Both mean GVI, sky GVI, middle GVI, and ground GVI in Harbin were higher than those in Changchun. SG was mainly driven by landscape patterns, and their explaining power decreased from intra-city to inter-city level. In Harbin, the major driving factors for GVI were patch density (PD), tree height, and total green space area (TA), and in Changchun, tree height (TH), edge density (ED), and aggregation index were the main ones. Pooled two cities' data showed that GVI was affected by TH, TA, and ED. Our findings highlighted that SG could be regulated by landscape configuration and large tree conservation, rather than species richness. Conclusions Utilization of internet big data, field survey, and remote sensing could provide a new basis for urban green infrastructure management from SG regulation, and our data is an example for this.

Context Internet-based street-view greenery (SG) is a new tool for evaluating urban green infrastructure, with proved vital services for residents, while until now, no report is on SG-aimed management from forest structure and landscape patterns. Objectives To find out plant composition, tree size, and landscape pattern's contribution to inter-and intra-city SG variations, and implication for SG-maximization management. Methods The SG was quantified by upper green view index (sky GVI), middle GVI, and ground GVI by using Baidu Street View in Harbin and Changchun; forest structure and landscape metrics were also measured in field survey, street-view picture, and remote sensing, then matched to 2 km x 2 km grids. Sampling density was 0.25-3 plots/km(2), securing uncertainty < 10%. Results Both mean GVI, sky GVI, middle GVI, and ground GVI in Harbin were higher than those in Changchun. SG was mainly driven by landscape patterns, and their explaining power decreased from intra-city to inter-city level. In Harbin, the major driving factors for GVI were patch density (PD), tree height, and total green space area (TA), and in Changchun, tree height (TH), edge density (ED), and aggregation index were the main ones. Pooled two cities' data showed that GVI was affected by TH, TA, and ED. Our findings highlighted that SG could be regulated by landscape configuration and large tree conservation, rather than species richness. Conclusions Utilization of internet big data, field survey, and remote sensing could provide a new basis for urban green infrastructure management from SG regulation, and our data is an example for this.