The research team, led by Jianhua Guo of the International Research Center of Big Data for Sustainable Development Goals and the Aerospace Information Research Institute of the Chinese Academy of Sciences, together with colleagues including Xiao Xiang Zhu of the Technical University of Munich, set out to address a stubborn gap in urban mental health research. Lab studies and small surveys have long suggested that exposure to greenery improves mood, but evidence linking tree cover to real-time emotional expression at the scale of entire nations has remained scarce. Social media, for all its noise, offers a rare window into how millions of people feel as they move through their daily environments.

Negative emotions are particularly important to track, the authors argue, because they spread rapidly through online networks. Previous work has shown that negative online news is shared more readily than positive content, and that emotional contagion dynamics amplified distress during the pandemic. If the built environment can dampen negativity at its source, urban forestry becomes not just an aesthetic or climate policy, but a public mental health intervention. The new study provides some of the most granular evidence yet for that proposition, quantifying the sentiment effects of canopy cover with a fixed-effects statistical framework that controls for meteorological, environmental and socioeconomic confounders.

The technical backbone of the analysis is a pair of deep-learning pipelines. First, the team mapped urban tree canopy across Brazil using high-resolution satellite images, building on their earlier nationwide canopy mapping work published in 2023 and their continent-wide South American assessment in 2024. Second, they harvested geotagged posts from Twitter, Instagram and Foursquare, applying multilingual sentiment analysis to translate each post into a sentiment score. The sentiment data drew on established approaches, including the multilingual universal sentence encoder for semantic retrieval and distant-supervision classifiers originally developed for Twitter sentiment classification. Because the posts carry location and time stamps, they could be aggregated to neighborhoods and days, then matched against tree canopy measured within 500-meter buffers around posting locations.

The headline result is strikingly precise: each 1-percentage-point increase in tree coverage was associated with a reduction of 0.028 in negative sentiment scores before the pandemic, and 0.042 during COVID-19, both statistically significant at P < 0.001. In other words, the association was nearly 1.5 times stronger during the pandemic than in the pre-pandemic period. The pattern aligns with prior findings that expressed sentiment shifted globally during COVID-19, and with studies showing that contact with blue-green spaces during lockdowns benefited mental health. But the new analysis goes further by measuring the specific ingredient of urban greenery that matters: trees themselves, mapped pixel by pixel from orbit.

Crucially, the relationship is not linear. The benefits of tree cover increased with canopy extent up to roughly 40 to 45 percent coverage, after which additional gains diminished. This saturating curve has immediate planning implications. A city moving from sparse canopy to a substantial urban forest stands to gain the most sentiment benefit, while pushing coverage beyond the saturation threshold yields progressively smaller emotional returns. The authors present these thresholds as evidence-based tree planning targets for urban mental health, framing the work as a contribution to Sustainable Development Goal 3.4, which concerns promoting mental health and well-being, and SDG 11.7, which calls for universal access to safe, inclusive and accessible green public spaces.

The team also probed how the sentiment benefits of trees vary across urban and climatic contexts, and the heterogeneity is instructive. The protective association was stronger in wealthier regions, echoing long-standing concerns about environmental inequity: affluent neighborhoods already tend to enjoy more canopy, and the analysis suggests they also extract more psychological value from it. Conversely, the benefits weakened under high temperature, high precipitation and high population density. Extreme heat and heavy rain may suppress the outdoor activity through which people actually experience trees, while dense crowds can dilute the restorative, stress-reducing qualities that quieter green settings provide. The findings resonate with earlier research linking green space to lower stress in deprived communities and with meta-analyses showing that tree traits cool urban heat islands, suggesting that trees’ mood benefits may be entangled with the thermal and social conditions in which they are encountered.

The fixed-effects modeling design deserves emphasis because it strengthens the causal interpretation. By comparing the same cities over time and across locations within them, while controlling for weather variables drawn from the ERA5 global reanalysis, impervious surface extent, nighttime lights as a proxy for economic activity, population density and municipal GDP per capita, the models strip away many alternative explanations for why greener places might appear happier. The team further bolstered the analysis with nonlinear curve-fitting techniques to characterize the dose-response relationship and heterogeneity analyses spanning climate zones and urban environments. All code is publicly available on GitHub, and the processed sentiment and control-variable datasets are archived on Figshare, allowing independent researchers to replicate and extend the results.

Still, the authors and observers of this literature caution about limits. Sentiment analysis of social media captures expressed emotion among platform users, not clinical diagnoses, and posting populations skew toward certain demographics. The study identifies an association, and while fixed effects and rich controls reduce confounding, they cannot fully rule out reverse causation or unobserved factors. Nonetheless, the consistency of the signal, its dose-response structure and its amplification during a documented mental health crisis make a compelling case that urban trees are doing measurable emotional work. The study joins a growing body of remote-sensing-based health research from the same group, including assessments of how residential greenness relates to long-term depression and anxiety risk, and how biodiversity-rich recreational areas near cities can serve as nature-based mental health solutions.

For city planners, the message is concrete. Trees are not merely decoration; they are infrastructure for psychological resilience, and their value becomes most apparent precisely when societies are under the greatest stress. The Brazilian dataset, spanning tropical metropolises and mid-sized cities over a five-year window that includes the pandemic, suggests a quantifiable planning rule: target canopy coverage in the 40 to 45 percent range where feasible, prioritize greening in lower-income districts that currently enjoy neither the canopy nor its benefits, and recognize that in hot, wet or densely built environments, trees may need to be paired with other design choices to deliver their full emotional dividend. As urban populations grow and climate change intensifies both heat and hardship, the case for treating every planted street tree as a small investment in collective mood has never been better quantified.

Subject of Research: Association between urban tree canopy coverage and reduced negative sentiment in social media posts across Brazilian cities, with stronger effects during COVID-19

Article Title: Urban trees associated with a reduction in negative sentiment with stronger links during COVID-19

Article References: Guo, J., Guo, H., Wang, J., Hong, D., Kruspe, A., & Zhu, X. X. (2026). Urban trees associated with a reduction in negative sentiment with stronger links during COVID-19. Nature Cities. https://doi.org/10.1038/s44284-026-00519-8

Image Credits: AI Generated

DOI: 10.1038/s44284-026-00519-8

Keywords: urban trees, tree canopy, negative sentiment, social media, mental health, COVID-19, remote sensing, Brazil, urban green space, sentiment analysis, Nature Cities, sustainable development goals