Stormwater runoff mitigation is a valuable service of urban forests, but a challenging service to incentivize through policy since trees within urban forests vary in their runoff mitigation potential depending upon their form, structure, and configuration. It is logistically impossible to assess relevant attributes of each individual tree in an urban forest. Our research proposes a novel strategy to address this issue, using remote sensing to classify tree canopy based on ecohydrological characteristics.
The objectives of this study are: (1) develop a typology which groups trees by characteristics which differentially impact hydrology (ecohydrological landscape characteristics; ELCs), and (2) qualitatively rank the runoff mitigation potentials of each group.
Using the literature, we identified 29 ELCs and evaluated them using a decision tree to select those that could both be effectively extracted from available data and influenced by design or management, resulting in 14 ELCs selected for further analysis. We downloaded publicly-available remote sensing and geospatial datasets (lidar, orthoimagery, land cover) for Montgomery County, Maryland, USA. We used the datasets to delineate tree canopy and extract values which approximate ELCs. Cluster analysis was used to determine which ELCs co-occur and to classify trees into seven different urban forest types.
Results/Conclusions
The typology was developed within Montgomery County, Maryland. The ecohydrological landscape characteristics (ELCs) that were extracted included tree height, canopy openness, presence of understory, and surrounding ground covers. Tree size was a meaningful metric in typology development, along with the distance to the nearest tree. Most trees in the study area occurred in patches. Tree types were ranked from most hydrologically effective (1) to least hydrologically effective (7) based on data from the literature and from a companion study assessing the water balance of three of the urban forest types. The trees that were most hydrologically effective occurred as large patches of trees with a complex canopy structure.
We conclude that trees in urban areas can be divided into groups based on their ELCs and that these groups are meaningful for policy implementation. Importantly, a typology using remote sensing can increase accessibility for a greater number of local governments in the future. We discuss implications for policymakers seeking to increase the environmental benefits of urban forests by promoting and conserving high-value landscape types. This repeatable classification scheme will allow localities to utilize abundant data resources to optimize their use of natural resources to better manage stormwater.