Analyzing the causal factors of carbon stores in a subtropical urban forest

•We used path analysis to explore causal drivers of carbon storage in urban forests.•Tree cover and land use both had direct and indirect effects on carbon storage.•Species composition, but not diversity, significantly affected carbon storage.•Use of socio-ecological indicators improves carbon storage estimates.

Studies of forests and urban forest ecosystems have documented the various biophysical and socioeconomic correlates of carbon storage. Tree cover in particular is often used as a determinant of carbon storage for local and national level urban forest assessments. However, the relationships among variables describing the biophysical and socioeconomic environment and carbon are not simple statistical ones. Instead, there are complex interactions that can have either a unidirectional causal effects, or produce indirect effects through interactions with other ecosystem structure and landscape characteristics. Thus, understanding the direct and indirect effects of structure, composition, and landscape characteristics is key to quantifying ecosystem services. This study used field data from plots across an urban watershed, site-specific biomass equations, and structural equation modeling of urban forest structure and landscape variables to quantify the causal influences of tree cover, land use, stand density, species composition and diversity on carbon stores. Our path analysis shows that the effect of tree cover on carbon stores is not only direct but also indirect and influential through basal area and composition. Findings suggest that species composition, species diversity and land use have much more complex relationships than previously reported in the urban forest literature. The use of path analysis in these types of studies also presents a novel method to better analyze and quantify these direct and indirect effects on urban forest carbon stores. Findings have implications for urban forest ecosystem assessments that use tree cover as the sole metric for inferring ecosystem functions and services.