Impacts of urban growth on vegetation carbon sequestration - Urban Remote Sensing: Monitoring, Synthesis and Modeling in the Urban Environment - page 262

19.1 Introduction

land and the atmosphere owing to human urban development
will be generated (Section 19.6).

There is growing consensus that humans play an important role
in modifying the global carbon cycle through land-cover/land-
use conversions (e.g., deforestation) and the burning of fossil
fuels, both highly associated with urban development (Boyle and
Lavkulich, 1997; Nowak and Crane, 2002; Pouyat et al ., 2002;
Pataki et al ., 2006; Churkina, 2008; Grimm et al ., 2008; Churkina,
Brown and Keoleian, 2010). Urban lands, accounting for approx-
imately 3% of the total US land area in 1995, were estimated
on average to be 1.6% less effective in absorbing carbon dioxide
by green vegetation than their immediately adjacent non-urban
surroundings (Imhoff et al ., 2004). By contrast, an urban system
may release a thousand more units of carbon dioxide to the
atmosphere than a forest ecosystem of the same area, due to the
burning of fossil fuels to produce adequate amount of energy for
its sustainment (Odum, 1997).
The accounting of carbon balance in urban areas that con-
siders vegetation productivities as well as human energy uses,
however, remains largely incomplete at regional to global scales.
Although the conceptual framework of the integrated urban eco-
logical and socioeconomic analysis became mature by the end
of the last century (Grimm et al ., 2000; Pickett et al ., 2001),
quantitative assessment of carbon fluxes between urban areas
and the atmosphere has been limited to a few metropolitan cities
where either field measurements of vegetation productivities or
detailed information on energy consumptionwas available (Kaye,
McCulley and Burke, 2005; Pataki et al ., 2009). The large-extent
urban carbon accounting relies increasingly on the availability of
data concerning carbon fluxes or stocks observed and modeled
beyond the local scale.
Remote sensing technique offers a unique opportunity for the
investigation of carbon dynamics of terrestrial lands at regional to
global scales. Satellite imagery and aerial photos provide spatially
and temporally continuous data on vegetation properties, includ-
ing vegetation types, phenology, and biophysical characteristics
such as leaf area index (LAI; Turner, Ollinger and Kimball, 2004).
These data captured at the spatial resolution of a few meters to
many kilometers have been applied widely to productivity and
biomass estimation in natural vegetation and agricultural land
uses (Prince and Goward, 1995; Haboudane et al ., 2002; Zhao
et al ., 2009). Remote sensing also provides information on human
settlement patterns (Herold, Scepan and Clarke, 2002), which
may be used to infer carbon emissions from urban land uses,
energy consumption, and transportation.
In this chapter, I will review measures and estimation of
vegetation productivities, with emphasis given to light-use effi-
ciency (LUE) models that utilize reflectance data gathered with
optical remote sensing to estimate gross and net primary pro-
duction (Section 19.2). Then, I will present a case study using
the LUE approach to calculate changes in gross primary produc-
tion (GPP) in the US Census South Atlantic division between
1992 and 2001; these changes will be related to urban growth
identified based on changes in the US Census housing-unit den-
sity during the roughly same period (Sections 19.3 and 19.4).
Finally, I will discuss the impacts of urban growth on vegeta-
tion productivities and outlook for future work (Section 19.5).
By integrating the LUE-based vegetation productivity estimates
with emission data prepared using energy and transportation
surveys, a comprehensive view of net carbon exchange between

19.2 Vegetation
productivities and
estimation

19.2.1 Vegetation productivities

The Earth consists of several massive carbon reservoirs including
the atmosphere, plants and other living organisms, soils, marine
sediments, and sedimentary rocks etc., where the element carbon
takes different forms (e.g., CO 2 and carbonate) as it cycles through
those reservoirs. The exchange of carbon between the atmosphere
and terrestrial ecosystems (including plants, animals, soils, and
soil microbial species) relies significantly on green vegetation
photosynthesis, the process by which CO 2 is transformed to
carbohydrate in the presence of solar radiation; and respiration,
the process by which CO 2 is released back to the atmosphere.
The total amount of carbon that enters an ecosystem through
photosynthesis during a certain time period, usually a year, is
referred to as gross primary production (GPP; Chapin, Matson
and Mooney, 2002). GPP indicates the maximum potential
carbon uptake from the atmosphere by plants in an ecosystem.
By subtracting the total plant respiration from GPP over the
same time span, net primary production (NPP) is obtained. This
is the net carbon obtained by plants through photosynthesis.
GPP and NPP are indicators for carbon flows between vegetation
and the atmosphere regardless of the presence of other living
organisms or soil carbon processes; therefore, they are referred
to as vegetation productivities in this chapter.

19.2.2 Estimation of vegetation
productivities

The direct measurement of GPP and NPP proves difficult for
large areas; therefore, alternative solutions rely onmodeling these
vegetation productivities at regional to global scales (Cramer
et al ., 1999; Zheng, Prince and Wright, 2003). For example,
Running and Hunt (1993) developed the Biome-BGC model
that simulates photosynthesis, respiration, and other ecosystem
processes for areas up to hundreds of square kilometers. This
model, combined with Landsat-based land-cover classification
and leaf area index as well as meteorological variables measured
at eddy covariance towers, was used to estimate GPP in two
hardwood and boreal forest sites in the United States (Turner
et al ., 2003).
The light-use efficiency (LUE) approach, also known as pro-
duction efficiency model, is a type of productivity models that
relies extensively on remotely-sensed biophysical characteristics
of land. The concept of the LUE approach resides in the corre-
lation between optical properties of green leaves and the leaf's
physiological functions (Tucker and Sellers, 1986). Green leaves
are highly efficient in absorbing shortwave radiation between
400 and 700 nm, due mainly to the presence of chlorophylls in