Impacts of urban growth on vegetation carbon sequestration - Urban Remote Sensing: Monitoring, Synthesis and Modeling in the Urban Environment - page 270

exurban development in South Atlantic division in 1992 and
2001. High-density urban and suburban development, compared
to rural or low-density exurban development, was showed to
reduce GPP by over 50% and approximately 20%, respectively,
in both years. These numbers were higher than previous finding
in four of the eight states located in my study area, where NPP
(usually half the amount of GPP) of the urban land-cover class
was estimated to be approximately 18% less than other natural
vegetation combined (Milesi et al ., 2003). The discrepancy may
be introduced by different definition of ''urban.'' In Milesi
et al . (2003), the authors defined urban as a type of land cover
based on remotely sensed nighttime light illumination, whereas
in this study urban density was determined based on Census
housing-unit density. Previous research (Imhoff et al ., 2000)
indicated that the urban land cover identified from night-time
illumination imagery may include both urban and suburban
densities as defined in this study.
The South Atlantic division for the region as a whole was
found to be more productive in 2001 as measured by GPP, which
increased by 1.70 g C m − 2 day − 1 comparing to the estimate
in 1992 (4.11 g C m − 2 day − 1 ). The monthly average of daily
minimum temperature was not statistically different between
1992 and 2001 according to the paired-samples t test, which
indicates minimum impacts of temperature on GPP. The region-
wide increase in vegetation productivity in the latter year may
partially resulted from the decline of incident solar radiation in
the former year following the volcanic eruptionofMt. Pinatubo in
June, 1991 (Ramachandran et al ., 2000; Tucker et al ., 2001; Zhao,
Brown and Bergen, 2007). Although radiometric effects of this
globally influential volcanic eruption was not explicitly corrected
before estimating GPP in this study, variation of changes in GPP
between 1992 and 2001 was still captured across different types
of urban growth.
Exurbanization, the conversion from rural to exurban densi-
ties, was found to enhance the GPP increment (i.e., higher than
the regional average increase of GPP) by 6.6%. Suburbanization,
the conversion from exurban to suburban densities, was found
to diminish the region-wide GPP increment (i.e., lower than
the regional average increase of GPP) by 15.2%. Urbanization,
the conversion from suburban to urban densities, was found to
minimize the region-wide GPP increment by 94.4%, where the
productivity gained through increased solar radiation in 2001
appeared to be nearly offset completely due to this high-density
settlement development. Urbanization reduced GPP by a greater
amount than the persistent urban densities, the latter reducing
the region-wide GPP increment by 91.9%. Both exurbaniza-
tion and suburbanization reduced GPP by a smaller amount,
compared to the corresponding persistent types, due possibly
to the remaining large proportions of vegetation cover in these
transitional categories than the persistent exurban and suburban
densities, respectively.
Spatial heterogeneity of changes in vegetation productivities
associated with urban and urban growth has been documented
in earlier research (Milesi et al ., 2003; Imhoff et al ., 2004; Zhao,
Brown and Bergen, 2007). Relationships between vegetation
productivities and human settlement patterns, especially those
patterns connected to andmeasured by demographic and socioe-
conomic transitions, however, are not well understood yet. The
case study in the Census South Atlantic division was among
the first attempts to explore those relationships by examining
urban growth at different intensities at local to regional scales.
This research approach links Census demographic changes and

landscape biophysical characteristics, and may be extended to
large geographic areas (such as the entire United States) where
demographic and vegetation data are publicly available. The
study presented in this chapter focused mainly on analysis of
GPP, whereas future work may be extended to the estimation of
NPP by incorporating respiration from ecosystemmodels such as
Biome-BGC (Running and Hunt, 1993). The modified approach
may also incorporate other environmental variables such as soil
moisture, soil nutrient availability, and vapor pressure deficit
(Hargrove and Hoffman, 2004; Hashimoto et al ., 2008) to simu-
late the real environmental condition for the analysis of vegetation
carbon exchange with the atmosphere.

Conclusions

In this study, urban growth and its impacts on vegetation pro-
ductivity measured by gross primary production (GPP) was
examined throughout the US Census South Atlantic division
between 1992 and 2001. In both years, GPP estimated based
on the light-use efficiency (LUE) approach was found to be the
highest in rural and low-density exurban areas. Between 1992
and 2001, GPP was estimated to increase by 1.70 g C m − 2 day − 1
or approximately 42% for the study area as a whole. Exurban-
ization, accounting for 8.53% of the study area, was showed
to be slightly less productive than the persistent rural densities,
slightly more productive than the persistent exurban densities,
and 6.6% more productive than the regional average. Subur-
banization and urbanization, accounting for 4.78% and 0.42%
of the study area, were associated with decline in GPP incre-
ment compared to the regional average. The dampening effects
of urban growth on vegetation photosynthesis, by the order of
decreasing strength, was found to be urban densities developed
from exurbs (E → U; 97.6% lower than the regional average
GPP increment), urbanization (94.4% lower), persistent urban
densities (U → U; 91.9% lower), suburbs developed from rural
densities (R → S; 37% lower), persistent suburbs (S → S; 33.8%
lower), and suburbanization (15.2% lower).
This study used publicly available Census demographic data,
Landsat-based land-cover/land-use data, vegetation greenness
index (i.e., AVHRR NDVI), climate variables on temperature
and solar radiation, and light-use efficiency parameters based on
empirical studies to identify urban growth and to estimate GPP.
Researchmethods may be applied to other geographic areas at the
comparable scale. The estimates of vegetation carbon fluxes may
be incorporated with the measured or modeled anthropogenic
CO 2 emissions from the burning of fossil fuels. Such integration
would contribute to better understanding of human carbon
impacts associated with cities and urban growth.

Acknowledgments

This work was supported by the 2008 FSU Council on Research
and Creativity First Year Assistant Professor summer research
program.TheGLDASclimatedatausedinthisstudywere
acquired as part of the mission of NASA's Earth Science Division,
and archived and distributed by the Goddard Earth Sciences
(GES) Data and Information Services Center (DISC). The author