| dc.description.abstract |
The urban heat island (UHI) effect is one of the many challenges facing urban development
as climate change and urbanization increase. The UHI effect refers to a phenomenon
in which dense concentrations of impermeable building materials absorb and trap solar
radiation. The gradual dissipation of heat raises ambient temperatures of urban areas
higher than the surrounding non-urban areas. It is anticipated that increases in ambient
temperatures and urban density will be a serious threat to public health as it increases
the risk of heat-related illnesses and mortalities.
Past studies have shown that increasing permeability and reducing density are key
to reducing the UHI effect. In urban developments, green spaces fulfill both of those
criteria. However, it is important to maximize the cooling effect (CE) and cooling
intensity (CI) of green spaces as space in urban developments are at a premium.
This study explored the factors for measuring urban density geospatially and identified
green space attributes that had the greatest impact on CE and CI in Los Angeles. The
green spaces analyzed were predominately parks managed by the City of Los Angeles.
The analysis was divided into three primary parts and utilized remote sensing and
geospatial information systems (GIS) techniques. First, the land surface temperature
(LST), was calculated using raster images captured by the Landsat-8 satellite. Using
the product of a thermal imaging satellite allowed for accurate temperature measurements
between all of the parks with no temporal variation in between them. Parks were then
overlaid on the LST raster to calculate CE and CI. Second, multiple geospatial analysis
products were created as inputs for a wholistic land classification of the Los Angeles
urban area. The products included: semi-supervised raster classification, normalized
difference vegetation index (NDVI), city zoning data, and urban density ratio. The
Urban density is defined as a ratio of the total number of inhabitants living within
a well-defined footprint of a city. As this project requires analysis of specific
areas, the city was discretized into census tracts prior to calculating the urban
density. The population of the census tract then was divided by the urban density
of the same census tract to produce the Urban Density Ratio. The inclusion of population
in this analysis was to account for anthropological effects that are not immediately
reflected in solely measuring building density. However, the inclusion of population
created potentially specious results for densely built-up areas with low population
such as industrial areas; a correlation coefficient was calculated in an attempt to
highlight the limitation.
Finally, all of the products were combined to create the land classification output.
Using the output, categories of land classification were divided into developed and
undeveloped land. Then sub-categorized by intensity of development to create the groups:
high-intensity development, medium-intensity development, and low-density development.
For vegetation density, the groups dense vegetation, sparse vegetation, and open area.
According to the results of the comparison between LST and the land classification
output, areas that are medium intensity developed have a higher mean temperature than
high intensity areas; which is intuitively contradictory. This discrepancy can be
explained by the correlation coefficient for each classification. Certain high intensity
(heavy industrial) areas have higher LST than medium intensity areas; however, there
are other high intensity (downtown Los Angeles) areas that are much cooler than medium
intensity areas.
The results of the park analysis indicated that medium sized parks performed the best
surpassing large parks in both CE and CI. The presence of vegetation had a high correlation
with lower CE And CI; however, the type of vegetation had a low correlation. Other
results of the analysis found that parks cool best when they’re close to each other
and when they are located in low and medium intensity developed areas. Moreover, the
results of this project indicate a higher correlation between contiguity brakes in
developed areas and lower LST than any one specific attribute of a park.
|
|