Journal Article

Street-view vegetation, human mobility, and dengue fever: context-dependent risk in urban green spaces

Jing, Fengrui, Hu, Jianxiong, Xiao, Jianpeng, Hu, Jinjing, Ma, Wenjun, Kang, Min, Zhou, Shuli, Chen, Panru, He, Guanhao, Liu, Tao, Gao, Yanhui

2026

Overview

Background The relationship between urban vegetation and dengue risk remains unclear, partly because most studies rely on static residential exposure and coarse greenness measures. This study examined how street-view vegetation and human mobility jointly influence neighborhood-level dengue risk in Guangzhou, China. Methods We analyzed 1,054 communities in Guangzhou using 3,972 locally acquired dengue cases from 2015 to 2019. Street-view images were used to derive grass, plant, and tree cover, while Landsat imagery provided NDVI. Row-standardized origin-destination mobility networks were used to construct mobility-weighted vegetation metrics. Negative binomial regression models with a population offset were fitted to estimate associations of local and mobility-weighted vegetation with dengue risk. Sub-group analyses were conducted by sex and age, and sensitivity analyses tested the robustness of the findings. Results In the local-only model, plant cover was significantly associated with lower dengue risk (IRR = 0.883, 95% CI: 0.807–0.967). In the full model, local grass (IRR = 0.885, 95% CI: 0.793–0.988) and plant cover (IRR = 0.807, 95% CI: 0.731–0.891) remained protective, while mobility-weighted tree exposure was strongly associated with higher dengue risk (IRR = 2.443, 95% CI: 2.171–2.749). Mobility-weighted grass (IRR = 0.836, 95% CI: 0.729–0.959) and plant cover (IRR = 0.597, 95% CI: 0.524–0.679) were protective, whereas NDVI was not significant in either local or weighted models. Model fit improved substantially after adding mobility-weighted vegetation metrics (AIC: 3859.14 vs. 4147.97). Associations were more evident among people aged 60 years or younger, while no greenery metric was statistically significant among older adults. Conclusions Urban dengue risk is shaped by both local streetscape vegetation and mobility-mediated exposure to vegetation in connected destination communities. Dense tree canopies in daily activity destinations may increase dengue risk, whereas understory plants and grass may be protective. These findings support mobility-aware urban planning and vector-control strategies.

• Publisher: BioMed Central,Springer Nature B.V,BMC
• Subject: Adolescent / Adult / Adults / Biostatistics / China - epidemiology / Climate change / Dengue / Dengue - epidemiology / Dengue fever / Disease transmission / Environmental conditions / Environmental Health / Epidemiology / Exposure / Female / Grasses / Green infrastructure / Greenspace / Guangzhou / Human mobility / Humans / Image acquisition / Landsat / Male / Medical research / Medicine / Medicine & Public Health / Middle Aged / Mobility / Mosquitoes / Neighborhood Characteristics - statistics & numerical data / Neighborhoods / Older people / Open spaces / Plants / Population Dynamics - statistics & numerical data / Public Health / Regression analysis / Regression models / Remote sensing / Residence Characteristics - statistics & numerical data / Risk / Roads & highways / Satellite imagery / Sensitivity analysis / Socioeconomic factors / Statistical analysis / Street view images / Streets / Understory / Urban planning / Vaccine / Vector-borne diseases / Vectors (Biology) / Vegetation