Original Analysis of Urban Surface Temperature Using Remote Sensing and Geographic Information System (GIS)

The study analysed variation in surface temperature (ST) in Makurdi Urban Area (MUA), Northcentral Nigeria. A total of 12 Landsat TM/ETM+ images were acquired in January, April and June of 1991, 1996, 2001 and 2006. The ST was estimated from the 12 Landsat TM/ETM+ images, grouped into seven classes, and the area of each ST class was determined using remote sensing and Geographic Information System (GIS). The ST magnitudes vary spatially from 27.5 o C (water bodies) to 50.7 o C (built-up land), representing an intensity of 23.2 o C. The mean seasonal ST varies from 32.4 o C-34.5 o C (cool-dry season), 35.5 o C-38.8 o C (hot-dry season) and 30.8 o C-31.4 o C (hot-wet season). The mean annual ST has increased from 32.9 o C (1991) to 35.9 o C (2006) with ST intensity of 3.0 o C. The ST classes of 27 o C-29 o C and 33 o C-37 o C recorded the highest loss and gain in area of -126.5km 2 and 94.5km 2 whereas ST classes of 29 o C-33 o C and 41 o C-45 o C recorded the least and highest per centage change in area of 22% and 768%. The result showed decreasing and increasing trends in the areas of cooler and warmer surfaces, which are attributed to increase in anthropogen surface materials, with higher heat storage capacities, due to urbanisation.


Introduction
Surface temperature, also refered to as brightness temperature (Chen et al., 2006), is the temperature of radiatively active natural and anthropogenic surface materials that absorb and store radiant energy.
Surface temperature (ST), often refered to as land surface semperature (LST), is one of the key parameters controlling the physical, chemical and biological processes of the earth (Pu et al., 2006). The ST or LST forms an important climate variable that related to climate change (Kayet et al., 2016).Surface temperature is important in urban studies because it has considerable impact on urban thermal environment causing urban heat isalnds (UHIs) (Dissanayake et al., 2019) and controls surface heat and water exchange with the atmosphere (Gallo et al., 1993).
Most studies in literature have attributed ST and air temperature patterns in urban areas to changes in surface energy balance. Due to the decrease in vegetation cover and increase in concretised surfaces in most cities, more incoming radiation is converted into sensible heat flux (QH) rather than latent heat flux (QE) (i.e., higher Bowen ration) resulting to higher surface and air temperature in cities relative nonurbanised areas (Wong & Yu, 2005). Tapper et al. (1981) reported that in rural areas in Christchurch, New Zealand, 40% of the net radiation is used in evaporation (QE), 26% goes into sensible heat (QH) and 33% into storage (QG). By contrast, in the industrial/commercial area, no energy is used for evaporation (QE), 64% is converted to sensible heat (QH), and 36% goes into storage (ΔQS).
Remote sensing is a global application methodology for assessing thermal effect of cities even in regions where pairs of urban and rural temperature records are not available (Gallo et al., 1993). Remote sensing, in conjunction with Geographic Information System (GIS), has been widely applied in detecting land use/land cover change, the basis for the inadvertent climatic modification of cities (Weng, 2001), assessing the distribution characteristics of surface temperature and surface urban heat island (SUHI) (Weng, 2001;Weng et al., 2006) and investigating the relationship between surface temperature and land use/land cover (Zhang et al., 2004;Chen et al., 2006;Yuan & Bauer, 2007). Urban surface heating and temperature exhibit spatial, seasonal and temporal variability in terms of magnitude, intensity and area coverage. Understanding the variation of these characteristics, in the context of land use/land cover change due to urbanisation, is fundamental in urban heat island mitigation, urban planning and design.
The major objectives of the study are to: (1) retrieve and estimate ST ( o C) from Landsat EM/ETM+ Images, (2) categorise ST ( o C) in seven classes and (3) analyse the variation in the magnitude and intensity of ST, and areas of the ST classes. www.scholink.org/ojs/index.php/uspa Urban Studies and Public Administration Vol. 4, No. 4, 2021 18 Published by SCHOLINK INC.

Study Area
Makurdi Urban Area (MUA) is located between latitudes 7 o 35 ‫|‬ -7 o 53 ‫|‬ N and longitudes 8 o 24 ‫|‬ -8 o 42 ‫|‬ E in Benue State, Northcentral Nigeria, and covers a land area of 800km 2 (Figure 1). The projected population of the study area in 2020 is 460,000 people based on 3.0% growth rate per annum. The MUA is subdivided into eleven political divisions known as council wards. The council wards that cover the metropolitan area are Mission, Clark/Market, Wadata/Ankpa, North Bank I and Wailomayo. Fiidi, Modern Market and North Bank II constitue the suburban council wards and the rural council wards comprise Bar, Mbalagh and Agan (Figure 1). In MUA, 60%, 30% and 10% of the population live in metropolitan, suburban and rural areas (Tyubee, 2021). Like other tropical wet and dry (Aw) climate regions, the study area experienced three thermal seasons namely cool-dry (November-January), hot-dry (February-April) and hot-wet (May-October). The mean seasonal air temperature is 26 o C (cool-dry), 31 o C (hot-dry) and 28 o C (hot-wet) (Tyubee, 2006). Annual rainfall, ranging from 900 to 1,500mm, dominantly occurs between April and October reaching its peak in September.

Retrieval and Classification of ST
The retrieval of ST from Landsat TM/ETM+ images was carried out pixel by pixel. The ST of both Landsat TM and Landsat ETM+ data was retrieved from the Thermal Infrared (TIR) band (band 6) using the procedure of Chen, et al. (2006). The procedure for retrieving ST involved the conversion of the Digital Numbers (DNs) of band 6 to radiation luminance (RTM6). The radiation luminance was then converted to satellite brightness temperature in degrees Kelvin ( o K). Since the DNs were not first converted to black body temperature, correction for emissivity (ε) was not necessary.
After the retrieval and estimation of ST from each of the 12 Landsat ET/ETM+ images, the ST was then

Data Analysis
The ST magnitudes were estimated pixel by pixel and mapped for each of the 12 Landsat ET/ETM+ images. The seasonal and annual ST magnitudes were computed from the seven ST classes for each of the 12 ST maps using the expression: Where ε, is summation; x, is the midpoint of each ST class, and f, is the number of pixels for each ST class.
The annual ST magnitudes were then averaged for the three (3) seasons. Standard deviation (σ) was applied to investigate the spatial variation in ST in the study area. The standard deviation was computed for the 12 ST maps from the expression: Where ε, x and f remain as in equation 1.
The retrieval, estimation and classification of ST from the 12 Landsat TM/ETM+ images were carried out using ERDAS Imagine 8.6 and ArcGIS 9.2 software.

Variation in ST Magnitude and Intensity
The      Table 3).

Discussion
The The result of seasonal changes in area surface temperature has revealed that hot-dry season (April) has the highest mean surface temperature and intensity of 37.0 o C and 3.1 o C, followed by cool-dry season (January) (33.5 o C and 2.1 o C) and least by hot-wet season (June) (31.5 o C and 1.3 o C). Moreover, there is higher variability in the distribution of surface temperature in April followed by January and least in June.
The result of seasonal variation in magnitude and intensity of ST confirmed Jin (2004) and Yuan and Bauer (2007) findings.
The spatial pattern and seasonal distribution of ST conformed with the observed patterns of land use/land cover changes in the study area (Tyubee, 2021) and were related to seasonal variation in cloud cover, surface and soil moisture, and vegetation cover which affect surface's solar radiation receipt, absorption and storage capacity. It has been reported that the highest and lowest solar radiation in the study area occurred in April and June respectively (Ojo, 1977). In June, cloud cover, humidity and abundant vegetation cover attenuated incoming solar radiation and also converted surface heat into latent heat through evapotranspiration (Weng, Lu, & Liang, 2006 (Bornstein, 1968;McPherson, 1994;King & Grimmond, 1997;Mills, 2004) which result in higher ST.

Conclusion
The over natural cover materials, compared to the previous years. The concluded that more warm surfaces have emerged, relative to the cooler ones, during the study period, and the spatial distribution of areas of warmer and cooler surfaces followed the spatial pattern of urbanisation in the study area.