Environmental Engineering Reference
In-Depth Information
use/land cover, vegetation concentration, tree cover, or other
environmental characteristics. Second, remote sensing is used to
develop estimates of population over small-areas to facilitate the
comparison of environmental hazard (or amenity) with demo-
graphic character. Here, remotely sensed imagery is incorporated
into a dasymetric mapping or areal interpolation algorithm to
derive fine-resolution population estimates where none exists, as
in many developing countries. Alternatively, such an approach
can be used to refine census data in many industrialized nations
where spatial population data are readily available.
From the beginning of the academic interest in environmen-
tal justice, researchers have recognized that land use and land
cover may be used as explanatory variables in statistical mod-
els of environmental equity. Though it is perhaps a tautology,
researchers have employed independent variables that capture
industrial land use/land cover to predict the locations of facili-
ties that release air pollution and other hazards associated with
industrial activity. The motivation for including such a land use
variable goes to the early arguments about causation in envi-
ronmental equity - is a hazardous facility located in a particular
location because of purely demographic reasons, or because there
is land available that suits the purpose of the facility (Been, 1994;
Boer
et al
., 1997)? Land use/cover data products derived from
remotely sensed data, such as industrial land cover data extracted
from land cover maps derived from Landsat imagery, have been
used in models of environmental equity (Mennis, 2005).
Another aspect of the use of remote sensing in environmental
justice research concerns the measurement of environmental
amenities. While much of the environmental justice literature
has focused on technological hazards associated with landfills, air
pollution, and toxic byproducts of industrial development, more
recently researchers have begun to focus on the relationship of
socioeconomic status with positive environmental characteristics
(Boone, 2008). One of the primary environmental amenities upon
which researchers have focused is green space, broadly defined,
i.e. open space, parkland, and vegetation, particularly in urban
areas, where access to non-residential, vegetated landscapes are
valued for recreation, health, and general well-being (Comber,
Brunsdon andGreen, 2008; Geoghegan, 2002; Strife andDowney,
2009). Because remote sensing can play a key role in quantifying
vegetation and identifying open space in urban environments
(Small, 2001), researchers have utilized remotely sensed imagery
to capture this type of environmental amenity for environmental
justice studies.
Several studies linking socioeconomic status to vegetation
have focused on assessment of quality of life, and have employed
the Normalized Difference Vegetation Index (NDVI) as a mea-
sure of vegetation character (Tucker, 1979). The NDVI exploits
the nature of healthy green vegetation to reflect relatively strongly
in the near-infrared wavelengths (NIR) and weakly in the visible
wavelengths (VIS), and can be considered a general measure
of the amount or density of green vegetation, though it may
also reflect other characteristics such as soil moisture content.
It is calculated as the ratio (NIR-VIS)/(NIR
with high income, high educational attainment, and other indi-
cators of socioeconomic advantage, as well as low population
density. The authors concluded that NDVI could be used to
capture a combined quality-of-life indicator that incorporates
both biophysical and demographic characteristics.
In a similar study of the Denver, Colorado region, Men-
nis (2006) found analogous relationships between NDVI and
US Census-derived socioeconomic characteristics in residential
land. Here, higher NDVI was associated with several indicators of
socioeconomic advantage, including higher educational attain-
ment and lower population density, as well as a lower percentage
of minority (i.e., non-white or Hispanic) residents. However,
the relationship between vegetation and socioeconomic status
was largely driven by certain types of developed land, includ-
ing wealthy, older neighborhoods with mature vegetation as
well as newer subdivision style developments with large lots
and well-maintained lawns. The author notes a likely feedback
effect between vegetation and socioeconomic status, such that
concentrated vegetation raises housing values, thus prohibiting
poorer segments of the population from buying into vegetated
communities. At the same time, wealthier people are better able
to create and maintain vegetated landscapes in a semiarid desert
environment such as Denver.
NDVI is not the only indicator of vegetation used in envi-
ronmental justice research. Li and Weng (2007) used spectral
mixture analysis (SMA) to extract green vegetation and impervi-
ous land covers from a Landsat Enhanced Thematic Mapper Plus
(ETM
+
) image to model quality of life in a study of Indianapolis,
Indiana. Green vegetation was positively correlated with sev-
eral US Census variables indicating socioeconomic advantage,
including educational attainment, employment, and housing
value, and was negatively correlated with poverty rate. Greater
impervious surface, on the other hand, was associated with lower
socioeconomic status.
Zhang, Tarrant and Green (2008) note that relationships
between socioeconomic status and open space may vary among
rural, suburban, and urban regions. In an analysis of Georgia, they
stratify US Census block groups into rural, suburban, and urban
regions and examine the relationship of a host of socioeconomic
variables with proximity to federally managed open space land in
rural areas. For quantifying open space in urban and suburban
areas the authors use vegetation data contained within the 1992
National Land Cover Data (NLCD) data set (Vogelman
et al
.,
2001), a digital land cover map derived from classification of
Landsat imagery. Results indicate that Georgia residents living
in close proximity to publicly managed land in rural areas, or
in areas with high vegetation concentration in suburban and
urban areas, are more likely to be wealthy, white, and have higher
educational attainment.
In an analysis of Terre Haute, Indiana, Jensen
et al
. (2004) use
Leaf Area Index (LAI) as a measure of vegetation. LAI is a ratio
measure of the area of ground covered by leaves. LAI was captured
in situ at points throughout the city as derived from below- and
above-forest canopy measurements of photosynthetically active
radiation (PAR). A continuous image of LAI for the entire study
area was derived by creating an artificial neural network (ANN)
model of LAI from green, red, and near-infrared reflectance
values contained in an Advanced Spaceborne Thermal Emission
Radiometer (ASTER) image. LAI and population density were
used to model household income and housing value. These
authors found that higher LAI and lower population density are
associated with higher income and higher housing values.
+
VIS), where higher
values indicate higher vegetation concentration.
Lo and Faber (1997) combined NDVI derived from Landsat
Thematic Mapper (TM) imagery with socioeconomic US Census
data for Athens-Clarke County, Georgia. Using principle compo-
nents analysis (PCA), these authors found that single dimensions
of variation within the data captured associations between NDVI
and socioeconomic characteristics. High NDVI was associated
Search WWH ::
Custom Search