Satisfaction with urban trees associates with tree canopy cover and tree visibility around the home

Urban trees support sustainability by providing environmental, economic, and social benefits, yet tree populations are under pressure and unevenly distributed, leading to inequities. Understanding how community perceptions—specifically satisfaction, defined as the discrepancy between expectations and experiences—relate to tree abundance and visibility near the home can inform equitable and effective urban tree policies. Prior work showed weak city-scale associations between canopy cover and satisfaction, and emerging eye-level greenness measures may better reflect lived experience than top-down metrics. The study asked: (1) Is satisfaction with urban trees associated with neighborhood greenness measures (NDVI, canopy cover, VGVI)? (2) Is satisfaction with tree management associated with these measures? (3) Do associations vary by neighborhood scale (100, 300, 500 m)? The authors hypothesized stronger positive relationships for VGVI than NDVI or canopy cover, and stronger associations at larger spatial buffers reflecting daily mobility and potential spatial aggregation effects.

The paper reviews research linking urban trees to health and wellbeing and notes inequities in tree distribution related to historic discrimination. Traditional management focuses on biophysical indicators, while social measures often miss perception responses such as satisfaction. Prior studies show mixed or weak links between top-down greenness (e.g., NDVI, canopy) and perceptions at broad scales, and that preferences can depend on biodiversity and presence/absence of trees. Eye-level measures (e.g., GVI from street view, VGVI from elevation models) align better with human visual experience and have been associated with mental health and wellbeing, but their relation to satisfaction with trees had not been tested. The authors highlight methodological challenges including metric choice (NDVI vs tree-specific metrics) and scale (modifiable areal unit problem), motivating a multiscale, perception-focused approach.

Ethics: Approved by University of Toronto Ethics Review Board (Protocol No. 00040945); informed consent obtained. Theoretical framework: Cognitive hierarchy model; satisfaction conceptualized as a specific, less abstract perception response closely linked to daily tree experiences. Context: City of Toronto, Canada, diverse urban form and population; municipal targets to increase canopy from 28% to 40%. Survey: Systematic, random, probabilistic online panel (Asking Canadians) capturing 223 respondents with postal codes. Measured: satisfaction with urban trees (8-item, 5-point scale), satisfaction with tree management (8-item), nature relatedness (NR-6), self-assessed tree knowledge, presence of a tree in front of home, environmental organization membership, demographics (education, age, tenure in neighborhood, home ownership, Canadian-born, ESL, ethnicity, gender). Greenness assessment: Computed three measures within Euclidean buffers (100, 300, 500 m) around respondents’ postal code areas. NDVI: Sentinel-2 (10 m) summer 2020 composite (cloud <10%), RED and NIR bands; values <0 removed; mean NDVI per buffer. Canopy cover: 2 m land cover (City of Toronto Automated Land Cover Analysis-2018), extracting tree class to compute percent canopy per buffer. VGVI (trees only): Eye-level tree visibility using Ontario 2 m digital surface model plus canopy layer; 300 m viewing distance, sampled every 20 m along streets (OpenStreetMap) within each buffer; computed with GVI R package, averaged per postal code. Multiple buffer distances used to assess MAUP effects and better represent daily experiences (100 m immediate vicinity; 500 m walkable neighborhood). Data analysis: R 4.2.1; confirmatory factor analysis (lavaan) and reliability (psych) for scales; generalized linear models (Gaussian) predicting satisfaction with trees and with management from each greenness metric at each buffer (18 GLMs), controlling for nature relatedness, tree knowledge, tree in front of home, years in neighborhood, age (median), Canadian-born, ESL, home ownership, university degree, white ethnicity, female gender, environmental organization membership. Assumptions checked via residual diagnostics and VIFs. Effects reported with 95% CIs and p-values.

Sample characteristics (n=223; 300 m buffer descriptive greenness): satisfaction with trees M=3.68, SD=0.72; satisfaction with management M=3.23, SD=0.84; NDVI M=0.36, SD=0.10; canopy cover M=0.26, SD=0.11; VGVI M=0.16, SD=0.07. Associations with satisfaction with urban trees (controlling covariates): - VGVI: positive and significant at all buffers; 500 m Estimate=2.50 (95% CI 0.93, 4.08; p<0.01; AIC 461.3), 300 m 2.46 (1.03, 3.88; p<0.001; AIC 459.9), 100 m 1.66 (0.27, 3.05; p=0.02; AIC 467.0). - Canopy cover: positive and significant at all buffers; 500 m 1.57 (0.61, 2.53; p<0.001; AIC 460.9), 300 m 1.70 (0.80, 2.59; p<0.001; AIC 458.1), 100 m 1.37 (0.51, 2.24; p<0.01; AIC 464.0). - NDVI: not significant; 500 m 0.92 (-0.16, 2.01; p=0.10), 300 m 0.97 (-0.04, 1.98; p=0.06), 100 m 0.79 (-0.20, 1.79; p=0.12). Associations with satisfaction with management: No significant associations with NDVI, canopy cover, or VGVI at any buffer (per Supplementary results). Scale effects: Associations for VGVI and canopy cover were stronger at larger buffers (300–500 m) than at 100 m.

Findings indicate that satisfaction with urban trees is more strongly related to tree-specific and eye-level visibility metrics (VGVI) than to generalized vegetation (NDVI), supporting the hypothesis that what people see at eye level better reflects their perceptions. Canopy cover, while tree-specific, remains a top-down 2D metric and may overestimate visible trees in dense built areas; VGVI likely better captures actual visual exposure. Stronger associations at 300–500 m align with daily mobility patterns and potential spatial aggregation (MAUP) effects, suggesting neighborhood-scale visibility and abundance shape satisfaction more than immediate vicinity alone. Lack of association with satisfaction in management suggests that managerial satisfaction reflects broader governance perceptions and expectations not directly tied to local greenness exposure. Results underscore the value of incorporating perception-based outcomes, particularly satisfaction with trees, into urban tree planning and monitoring to align enhancement efforts with community experiences and to address inequities in tree distribution.

The study demonstrates that residents’ satisfaction with urban trees is positively associated with neighborhood tree canopy cover and especially with eye-level tree visibility (VGVI), with stronger effects at larger neighborhood scales. NDVI was not associated, and greenness was not linked to satisfaction with tree management. These findings highlight the importance of using eye-level, tree-specific exposure metrics and considering neighborhood-scale experiences when planning and evaluating urban tree initiatives. Urban forest management should integrate satisfaction measures alongside biophysical metrics to better reflect community expectations and experiences. Future research should test these relationships across diverse contexts and cities, incorporate additional neighborhood characteristics and interaction effects, explore non-linear associations, and leverage larger datasets and refined analytical techniques.

Key limitations include a modest sample size with postal codes (n=223) and greenness data limited to the City of Toronto, restricting generalizability. The analysis did not incorporate neighborhood-level characteristics (e.g., socio-economic disadvantage), interaction effects (e.g., age groups), or urbanity gradients due to the individual-level modeling frame and data constraints. Exposure metrics were assessed with Euclidean buffers and linear models; other functional forms and mobility-aware exposures were not tested. The study relied on cross-sectional data and averaged greenness within buffers, and VGVI used street-only viewpoints and a 300 m viewing distance. Broader validation with larger, multi-city datasets, more covariates, non-linear models, and alternative exposure definitions is warranted.