Current street tree communities reflect race-based housing policy and modern attempts to remedy environmental injustice

Abstract

Humans promote and inhibit other species on the urban landscape, shaping biodiversity patterns. Institutional racism may underlie the distribution of urban species by creating disproportionate resources in space and time. Here, we examine whether present-day street tree occupancy, diversity, and composition in Baltimore, MD, USA, neighborhoods reflect their 1937 classification into grades of loan risk-from most desirable (A = green) to least desirable (D = "redlined")-using racially discriminatory criteria. We find that neighborhoods that were redlined have consistently lower street tree α-diversity and are nine times less likely to have large (old) trees occupying a viable planting site. Simultaneously, redlined neighborhoods were locations of recent tree planting activities, with a high occupancy rate of small (young) trees. However, the community composition of these young trees exhibited lower species turnover and reordering across neighborhoods compared to those in higher grades, due to heavy reliance on a single tree species. Overall, while the negative effects of redlining remain detectable in present-day street tree communities, there are clear signs of recent investment. A strategy of planting diverse tree cohorts paired with investments in site rehabilitation and maintenance may be necessary if cities wish to overcome ecological feedbacks associated with legacies of environmental injustice.

Keywords: community composition; ecosystem services; environmental justice; housing policy; street trees; tree canopy; urban biodiversity.

FIGURE 1

(a) Historic HOLC grade designations for 36 neighborhoods across Baltimore City, MD (I). We subsampled these neighborhoods (II) using equal‐area 3.3 ha hexagonal polygons (II) that contained both living street trees and viable locations where trees could be planted that were currently dead or empty (black dots in III and IV satellite imagery). (b) Occupancy of a given street tree location varied among HOLC grades and by tree size class. Each gray datapoint is the proportion of total viable street trees within a 3.3 ha hexagon that currently contains large (II), medium (III), or small trees (IV) or are currently empty without a living tree present (I). (c) Neighborhoods that were redlined (grade D) consistently contain a less diverse community of living street tree species as shown by individual‐based accumulation curves using three diversity metrics that range in weighting the importance of species number versus relative abundance (V). Note: The analysis only includes street trees; any park trees (i.e. the top right of (a)III and (a)IV) were removed.

FIGURE 2

Tree size distribution, community composition, reordering, and turnover of tree communities by neighborhood HOLC grade. (a) Smoothed kernel density estimate of tree sizes (dbh) for each HOLC grade. The dotted lines indicate the dbh cutoffs used to classify “small” and “large” tree designations used for the “old” versus “young” community comparisons. (b) Nonmetric multidimensional scaling (NMDS) of large and small tree communities. Ellipse denotes the 95% confidence interval for that distribution of points. (c) Reordering and species turnover of large and small trees among HOLC grades. Significant differences (p < 0.05) are denoted by letters.

FIGURE 3

Rank abundance curves of large and small trees for each HOLC grade. The five most abundant species are labeled. Neighborhoods that were redlined have a less even (steeper curve) community structure of young (small) trees dominated by Acer rubrum.