Yards, block groups, and vegetation cover measures

Residential yards are a significant component of urban socio-ecological systems; residential land covers 11% of the United States and is often the dominant land use within urban areas. Residential yards also play an important role in the sustainability of urban socio-ecological systems, affecting biogeochemical cycles, water, and the climate via individual- and household-level behaviors. Vegetation, such as trees and grasses, are unevenly distributed across front and back yards, and we sought to understand how similar yards are to each other when compared to their neighboring yards and neighborhoods using aerial imagery. There are many ways to measure yard similarity, and we compared several measures to account for different definitions of ‘neighborness’. We examined the spatial autocorrelation of several yard vegetation characteristics in both front and backyards in Boston, MA, USA. Our study area included 1,027 Census block groups (sub-neighborhood areas) and 175,576 parcels with matched front-backyard pairings (n = 351,152 yards in total) across Boston’s metropolitan area. This data package contains 1) 351,152 yard spatially-referenced yard polygons with five measures of vegetation summarized, 2) the containing block groups, and 3) and *.R script that replicates the analyses reported in Locke, D. H., Ossola, A., Minor, E., & Lin, B. B. (2021). Spatial contagion structures urban vegetation from parcel to landscape. People and Nature, 00, 1–15. https://doi.org/10.1002/pan3.10254

Methods

1. Study Area

This study focused on the Boston, MA, metropolitan region (42°21′29″N 71°03′49″W), an area of approximately 703 km2. The region has a humid continental climate (mean annual temperature = 9.6 °C; mean annual precipitation = 1233 mm) (PRISM Climate Group 2015) and was historically covered with mesic forests. Forty-four percent of the land area is residential (Ossola et al., 2019a), which is consistent with other urban areas in western countries such as Baltimore, MD (Avolio et al., 2020), Chicago, IL (Lewis et al., 2019), Adelaide, (Australia)(Ossola et al., 2021), Edinburgh (Scotland), Belfast (Northern Ireland), Cardiff (Wales), and Leicester and Oxford (England) (Loram et al., 2007), and represents more than twice as much land area as parks and open spaces (18.43%) (Ossola et al., 2019b). Backyards compose 14% of all urban land area and contain ~21% of all tree canopy cover; front yards cover ~8% of the area and have ~8% of the study area’s tree canopy cover (Ossola et al., 2019b).

2. Open Data

Classified LiDAR point cloud data (year 2014) were obtained from the US Geological Survey (“MA Post-Sandy CMPG 2013–14”, NPS = 0.7 m, vertical and horizontal accuracy = 0.05 m and 0.35 m, respectively). High-resolution RBG-NIR imagery (1 m ground resolution, year 2014) were obtained from the National Agriculture Imagery Program (NAIP, USDA). Residential parcel polygons, building footprints, and road centerline data were downloaded from the open data portals of the Commonwealth of Massachusetts (2017) and the City of Boston (2017).

3. Geospatial analyses

All front, corner, and backyards contained in all residential parcels with a house were located and classified in ArcGIS Desktop 10.5 (ESRI, Redlands, CA) by using the workflow described in Ossola and others (2019a, 2019b). Briefly, each house centroid was identified to fit an offset line perpendicular to the closest street centerline. Front and backyards were then located by splitting each parcel polygon with a dividing segment, perpendicular to the offset line, passing through the house centroid, and extending to the parcel’s border. Yards were classified by attributing the front yard as the closest unit to the respective road centerline. Corner yards, which lack clear front/back sides, were assigned to all parcels located within 15 m from street intersections and were excluded from analyses. The workflow used to locate and classify yards exceeded 98% accuracy (Ossola et al., 2019a). Vegetation maps detailing tree height, canopy volume, and tree and grass covers were modelled and validated for their accuracy based on the LiDAR and RBG-NIR imagery as detailed in previous papers (Ossola et al., 2019a, 2019b). Briefly, tree canopy height was extracted from a canopy height model (1.5 m ground resolution) interpolated from the LiDAR data in ArcGIS Desktop 10.5 (ESRI, Redlands, CA). Tree and grass covers were modelled at 1.5 m resolution by using maximum likelihood supervised classification of ~100,000 pixels manually attributed to one of three land cover classes (i.e., tree, grass and non-vegetated cover), and based on the tree canopy height map and the RGB-NIR imagery (Singh et al., 2012). The average vertical accuracy of the tree height data, as recorded by the LiDAR point cloud, is 5.3 cm. The accuracy of the grass and tree canopy cover classification is 91.7% and 98.9%, respectively (Ossola & Hopton, 2018a). Canopy volume was calculated as the product of tree canopy cover and height within each pixel, assuming this volume to be completely occupied by vegetation (Ossola & Hopton, 2018a, 2018b), which overestimates total volume. Because these remotely sensed data view the earth from above, and tree canopy overhangs turf, the turf estimates are plausibly underestimates (Akbari et al., 2003). 

References

Akbari, H., Rose, L. S., & Taha, H. (2003). Analyzing the land cover of an urban environment using high-resolution orthophotos. Landscape and Urban Planning, 63(1), 1–14. https://doi.org/10.1016/S0169-2046(02)00165-2

Avolio, M. L., Blanchette, A., Sonti, N. F., & Locke, D. H. (2020). Time Is Not Money: Income Is More Important Than Lifestage for Explaining Patterns of Residential Yard Plant Community Structure and Diversity in Baltimore. Frontiers in Ecology and Evolution, 8(April), 1–14. https://doi.org/10.3389/fevo.2020.00085

Lewis, A. D., Bouman, M. J., Winter, A. M., Hasle, E. A., Stotz, D. F., Johnston, M. K., Klinger, K. R., Rosenthal, A., & Czarnecki, C. A. (2019). Does nature need cities? Pollinators reveal a role for cities in wildlife conservation. Frontiers in Ecology and Evolution, 7(JUN), 1–8. https://doi.org/10.3389/fevo.2019.00220

Loram, A., Tratalos, J., Warren, P. H., & Gaston, K. J. (2007). Urban domestic gardens (X): The extent & structure of the resource in five major cities. Landscape Ecology, 22(4), 601–615. https://doi.org/10.1007/s10980-006-9051-9

Ossola, A., & Hopton, M. E. (2018a). Climate differentiates forest structure across a residential macrosystem. Science of the Total Environment, 639, 1164–1174. https://doi.org/10.1016/j.scitotenv.2018.05.237

Ossola, A., & Hopton, M. E. (2018b). Measuring urban tree loss dynamics across residential landscapes. Science of The Total Environment, 612, 940–949. https://doi.org/10.1016/j.scitotenv.2017.08.103

Ossola, A., Jenerette, G. D., McGrath, A., Chow, W., Hughes, L., & Leishman, M. R. (2021). Small vegetated patches greatly reduce urban surface temperature during a summer heatwave in Adelaide, Australia. Landscape and Urban Planning, 209. https://doi.org/10.1016/j.landurbplan.2021.104046

Ossola, A., Locke, D. H., Lin, B., & Minor, E. (2019a). Greening in style: Urban form, architecture and the structure of front and backyard vegetation. Landscape and Urban Planning, 185(November 2018), 141–157. https://doi.org/10.1016/j.landurbplan.2019.02.014

Ossola, A., Locke, D. H., Lin, B., & Minor, E. S. (2019b). Yards increase forest connectivity in urban landscapes. Landscape Ecology, 7(12). https://doi.org/10.1007/s10980-019-00923-7

Singh, K. K., Vogler, J. B., Shoemaker, D. A., & Meentemeyer, R. K. (2012). LiDAR-Landsat data fusion for large-area assessment of urban land cover: Balancing spatial resolution, data volume and mapping accuracy. ISPRS Journal of Photogrammetry and Remote Sensing, 74(November), 110–121. https://doi.org/10.1016/j.isprsjprs.2012.09.009