CHAPTER THREE

RESEARCH METHODOLOGY

3.1 1ntroduction

This chapter concern with data need, sources of data, instrument for data collection, method of
collection and analysis of data. Thus, it focuses on the research methodology and design that will
be adopted in carrying out this study on examination of the relationship between population
concentration and urban heat island in Awka, Anambra State.

3.2 Research Design

The research design constitutes the blueprint for collection, measurement and analysis of data;
thus, it provides an outline of what the researcher is going to do in terms of operational
implication and final analysis. The research design adopted for the purpose of this study is the
Experimental Design. This research involves the use of geospatial analysis of Landsat satellite
imageries to analyze the Land Surface Temperature (LST) in Awka, the use of ArcGIS and
population data to assess the spatiotemporal distribution of population in Awka, and The
Pearson’s Product Moment Correlation Analysis use to examine the relationship between
population concentration and urban heat island in Awka, Anambra State.

3.3 Data Need

The data required for this study include administrative Shapefile of Awka for boundary,
LANDSAT 30m resolution satellite data for Land Surface Temperature (LST) Estimation, and
100m resolution spatial time-series population data.

3.4 Sources of Data

3.4.1 Secondary Data

The secondary data for this study are from the Office of the Surveyor-General of The Federation
(OSGOF), The United States Geological Survey (USGS) Website, Global Human Settlement
Layer (GHSL) Website. However, administrative shapefile is sourced from the Office of the
Surveyor-General of The Federation (OSGOF), LANDSAT data acquired from the Earth
Explorer domiciled on the United States Geological Survey (USGS) Website, and Population
data is acquired from the Global Human Settlement Layer (GHSL) Website.

3.5 Instrument for Data Collection

3.5.1 Software

The software that was used for this study was for the processing and analysis of the datasets
required for the study. The software has a unique compatibility with the hardware devices, hence
enabling ease of access and transfer of data across the different platforms. They include;

1. ArcGIS 10.8.2 – This is a powerful program from ESRI that enables for the compilation,
analysis and distribution of geographic data; and managing geographic information in a
database. Creation of embellished maps is executed using this workspace. The population
analysis and LST estimation are carried out using this software.

2. Microsoft Excel and Word 2016: Microsoft word aided in the compilation of results and
the theoretical aspects of the study. Excel is used for the statistical analysis.

3.6 Method of Data Collection

3.6.1 Data Acquisition

The LANDSAT data were downloaded from USGS Earth Explorer. This study uses LANDSAT
satellite imagery (ETM+ and OLI sensors) to retrieve Land Surface Temperature (LST) over the
area for the three different periods (16 th December 2000, 18th December 2012, and 20th December
2023). However, the mono-window algorithm is applied to the thermal bands (band 6L for
ETM+ and bands 10 and 11 for OLI) to calculate LST. Furthermore, the population data is
disaggregated from administrative units to grid cells using the Global Human Settlement Layer
(GHSL). Meanwhile, this study uses residential population data from the CIESIN GPWv4.11
dataset, which provides estimates for 1975-2020 and projections to 2050 and 2030.

3.7 Method of Data Analysis

3.7.1 Objective 1
To analyze the Land Surface Temperature (LST) in Awka, requires Landsat 30m resolution data
for LST from 2000 to 2023. However, LST of Awka acquired from the earth explorer domiciled
on the United States Geological Survey (USGS) website

3.7.1.1 Land Surface Temperature (LST) Retrieval

The mono-window algorithm method is adopted to retrieve the LST from the imageries selected
for this study. The Landsat-5 Enhanced Thematic Mapper Plus (ETM+) band 6L (10.4 – 12.5
μm), and Operation Land Imager (OLI) 10 and 11 (10.60 -11.19 μm) have a spatial resolution of
30m each, which is considered suitable as shown by many works of literature for capturing the
multifaceted intra-urban temperature differences thus making it effective for climate analysis.
For the Landsat ETM+ sensor, images in the thermal band are taken twice: one in the low-gain
mode (band 6L) and the other in the high-gain mode (band 6H). Band 6L is used to image
surfaces with high brightness, whereas band 6H is for low brightness. Band 6L will be used in
this study, due to errors in the 6H band. Consequently, the LANDSAT thermal bands are used to
retrieve LST over the study area for the three different periods (2000, 2012, and 2023) based on
the following steps:

STEP 1. Conversion of Digital Numbers (DN) of the bands to Spectral Radiance The DN of
the thermal band of ETM+ are converted into spectral radiance values for each of the
investigated years using the following equation –

( )
LMAX −L
L λ= MIN
×(DN −1)+ LMIN −−−−−−−−−(1)
QCALMAX−Q CALMIN

Where;

L λ = The spectral radiance at the sensor’s aperture in W m−2 sr−1 μ m−1

L MAX = Spectral radiance scaled to QCALMAX in Wm−2 sr−1 μ m−1, available in the metadata.

L MIN= Spectral radiance scaled to QCALMIN in Wm−2 sr−1 μ m−1, available in the metadata.

QCALMAX = Maximum quantized calibrated pixel value (corresponding to L MAX ) in DN = 255

 QCALMIN = Minimum quantized calibrated pixel value (corresponding to L MIN) in DN = 1

DN = Digital Number of the Band

For Landsat 8 with file data on Radiance Multiplier (M) and Radiance Add (B), the thermal
infrared (TIR) band is converted into spectral radiance (Lλ) using the equation -

𝐿𝜆 = M L QCAL + A L_ _ _ _ _ (2)

Where;

Lλ = The spectral radiance at the sensor’s aperture inWm−2 sr−1 μ m−1

M L = Band Specific Multiplicative Rescaling factor, gotten from the metadata of the image

QCAL = Quantized and calibrated standard product pixel values (Digital Number

A L= Band Specific Additive Rescaling factor, gotten from the metadata of the image

STEP 2. Conversion from Spectral Radiance to At-Satellite Brightness Temperature Spectral

radiance values for the bands are then converted to radiant surface temperature under an
assumption of uniform emissivity using pre-launch calibration constants for the Landsat ETM+
sensor implemented into this equation;

𝑇=
k2
K1 − 273.15 _ _ _ _ _ _ (3)
ln( + 1)
Lλ

Where;

T = At-satellite brightness temperature in Degrees Celsius

Lλ = Spectral radiance in Wm−2 sr−1 μ m−1(gotten from equations 1 and 2)

K 1 = Band-specific thermal conversion constant from the metadata

( K 1_CONSTANT_BAND_X, x is the thermal band number)

K 2 = Band specific thermal conversion constant from the metadata

 ( K 2_CONSTANT_BAND_X, x is the thermal band number)

-273.15 = Constant for conversion from Kelvin to Degrees Celsius

STEP 3. Correcting for Land Surface Emissivity (LSE)

Normalized Differential Vegetative Index (NDVI) is used to estimate LSE for emissivity
correction. The temperature values obtained using Equation (3) are referenced to a blackbody.
Therefore, corrections for spectral emissivity (ε) become necessary according to the nature of
land cover. This equation was used;

e = 0.004Pv + 0.986 _ _ _ _ _ _ (4)

Where,

e = Land Surface Emissivity

0.004 and 0.986 = Constants for emissivity estimation

PV = Proportion of vegetation given by the equation –

𝑃𝑉 =(
2
NDVI −NDVI min
) _ _ _ _ _ (5)
NDVI max− NDVI min

NDVI = Normalized Differential Vegetation Index for each of the years

The index is defined by the equation below -

𝑁𝐷𝑉𝐼 =
NIR−RED
__ _ _ _ _ _ (6)
NIR+ RED

Where, NIR and RED are the reflectance in the near-infrared and red portion of the
electromagnetic spectrum respectively.

Thus, For LANDSAT 7 ETM+,

𝑁𝐷𝑉𝐼 =
BAND 4−BAND 3
_ _ _ _ _ _ (7)
BAND 4+ BAND 3
Where, Band 4 and Band 3 are the reflectance bands in the near-infrared and red portion of the
Electromagnetic spectrum of the Enhanced Thematic Mapper Plus respectively.

For LANDSAT 8 OLI,

𝑁𝐷𝑉𝐼 =
BAND 5−BAND 4
_ _ _ _ _ _ (8)
BAND 5−BAND 4

Where, Band 5 and Band 4 are the reflectance bands in the near-infrared and red portion of the
Electromagnetic spectrum of Operation Land Imager (OLI) sensor of Landsat 8 respectively.

NDVImin = Minimum value of NDVI for that year

NDVImax = Maximum value of NDVI for that year

STEP 4. Estimation of the Land Surface Temperature

Finally, after corrected the emissivity in equation (4), the LST is estimated using the equation;

𝐿𝑆𝑇 = × ln(ℇ)
BT BT
× __ _ _ _ (9)
1+ W P

Where,

LST= Land Surface Temperature in Degrees Celsius

BT = At-satellite brightness temperature

W = Wavelength of emitted radiance (µm)

𝑝 = H× (1.438 × 10−2𝑚 𝐾) = 14380 _ _ _ _ _ _ _ (10)

C
S

H = Planck’s constant (6.626 × 10−34𝐽𝑠)

S = Boltzmann constant (1.38 × 10−23𝐽/𝐾)

C = Velocity of light (2.998 × 108𝑚/𝑠)

e = LSE
3.6.2 Objective 2

To assess the spatiotemporal distribution of population in Awka during the study period requires
100m resolution spatial time series population data from the Global Human Settlement Layer
(GHSL) website.

3.7.2.1 Population Analysis

The preprocessed population data is clipped to the boundary of the study area, per year.
However, spatial raster dataset depicts the distribution of residential population, expressed as the
number of people per cell. Residential population estimates between 1975 and 2020 in 5years
intervals and projection to 2025 and 2030 derived from CIESIN GPWv4.11 were disaggregated
from official census data in administrative units to grid cells, informed by the distribution,
volume, and classification of built-up as mapped in the Global Human Settlement Layer (GHSL)
global layer per corresponding epoch.

3.7.3 Objective 3

To determine the relationship between population distribution and LST intensity in Awka,
Pearson’s product moment correlation Analysis is carried out.

3.7.3.1 Population/LST Relationship Examination

a. To determine the relationship between LST and population distribution in the study area, the
Band Collection statistical geoprocessing tool on ArcGIS is used. This tool provides statistics for
the multivariate analysis of a set of raster bands. When using the Compute Covariance and
Correlation Matrices option is enabled, the covariance and correlation matrices are output as well
as the basic statistical parameters, such as the values of minimum, maximum, mean, and standard
deviation for every layer. Correlation ranges from +1 to -1. A positive correlation indicates a
direct relationship between two layers, such as when the cell values of one layer increase, the cell
values of another layer are also likely to increase. A negative correlation means that one variable
changes inversely to the other. A correlation of zero means that two layers are independent of
one another.
b. To further confirm the results of (a), random points are generated and evenly distributed
across the study area, and values of both variables are extracted onto the points. These values are
used to develop a scatter plot of the values, to display the statistical relationship between both
variables using the Pearson’s Product Moment Correlation analysis. Pearson’s Product Moment
Correlation analysis is carried out to determine the relationship between population and land
surface temperature. The Pearson’s Product Moment Correlation Analysis is a statistical method
that tests the measures of linear association between two quantitative variables, with the linear
association going from +1 to -1 in decreasing order of strength. This equation is given as:

𝑟=
( ∑ X )( ∑ Y )
∑ XY −
N _ _ _ _ __ _ (11)
¿¿

Where r= Correlation Coefficient

X = Independent variable, which is the Population data

Y = Dependent variable, which is the land surface temperature readings

N = Number of observations

To test the strength of the correlation, the coefficient of determination was

used and given by:

C/D = r 2

Where, C/D = Coefficient of determination

2
r = Correlation coefficient