Urban forest biodiversity and cardiovascular disease: Potential health benefits from California’s street trees

Discussion In this section we first discuss our main results and compare our effect sizes to those in the lit- erature for greenspace and other interventions to prevent cardiovascular disease. Next, we explore possible mechanisms that might link biodiversity and health. Finally, to facilitate inter- preting our results, we simulate benefits from a tree planting project in the city of Los Angeles. We tested our null hypothesis that there was no association between biodiversity and mor- tality from cardiovascular disease and found that urban forest diversity and tree abundance are associated with lower mortality rates for Heart Disease and Stroke, and the effect was larger, more stable, and more robust for Heart Disease. Our results were insensitive to varying ele- ments of the empirical design. These results are consistent with studies that assess the relation- ship between neighborhood greenness and cardiovascular mortality [19–21] and all-cause mortality [27]. Unlike for biodiversity, we failed to find a strong relationship between exposure to trees per 100,000 individuals and mortality. When assessing changes in heart disease mortality in response to changes in greenspace exposure from Emerald Ash Borer infestations, Donovan et al. [17] found a robust relationship for lower respiratory disease mortality and weaker results for heart disease mortality. For heart disease, the effect was not statistically significant. We, too fail to find a relationship between the two, however our analysis differs in three important

ways. First, we are assessing a cross-sectional association whereas Donovan et al. [17] assesses a dynamic relationship within a causal framework. Second, our tree exposure variable is based on the number of trees sampled by our arborists, which may vary with the total number of trees per zip code, introducing measurement error. Finally, we include a measure of biodiver- sity, which is often but not always a function of the size of a forest stand. Given this relation- ship, measures of greenspace that do not include biodiversity may pick up spillover effects from biodiversity. Possible mechanisms In our conceptual framework, we discussed three hypotheses proposed by Aerts et al. [16] as to how biodiversity may affect cardiovascular disease: enhanced immune functioning, psycholog- ical restoration due to biophilia, and dilution of disease risk for vertebrates. Given that we find that biodiversity is associated with health benefits, we might expect humans to be more aware of benefits from psychological restoration as compared to those from enhanced immune func- tioning or dilution of disease risk. If so, we would expect that higher biodiversity would be cor- related with household income, as higher income households would have the means to invest in more restorative greenspace. S2 Fig plots binned scatter points for the Shannon Index against median 2016 household income. We fail to find a positive correlation between income and biodiversity, suggesting that, in this case, psychological restoration may not be the main

pathway by which biodiversity could affect cardiovascular disease. This figure and the second panel, which shows a positive relationship between the Shannon Index and pollution burden, also allay some concerns that selection bias in our arborist data could be driving our results. The Shannon Index appears uncorrelated with income, which explains why adding controls did not greatly affect the coefficient in column six of Table 3, and the Shannon Index appears positively correlated with pollution burden. To get a sense as to whether our results could be driven by the dilution hypothesis, we com- pared the county-level rate of Lyme disease per 100,000 person-years to the Shannon Index for 22 California counties with biodiversity data. S3 Fig in Supporting Information, a plot of binned scatter points, fails to show a negative relationship between the Shannon Index and PLOS ONE Biodiversity and cardiovascular disease PLOS ONE | https://doi.org/10.1371/journal.pone.0254973 November 3, 2021 13 / 23