Analysis of Land Surface Temperature due to Dynamics of Green Spaces and Water Bodies Using Geospatial Techniques in Gida Kiremu, Limu and Amuru Districts, Western Ethiopia

Analysis of the correlation between indices (Normalized Difference Vegetation Index, Normalized Difference Barren Index and Modified Normalized Difference Water Index) and land surface temperature is used to natural resources and environmental studies. This research aimed to analysis of Land Surface Temperature due to dynamics of Different Indices (NDVI, NDBaI and MNDWI) Using Remote Sensing Data in three selected districts (Gida Kiremu, Limu and Amuru), western Ethiopia. From thermal and multispectral bands of landsat imageries (Landsat TM of 1990, landsat ETM+ of 2003 and landsat OLI/TIRS of 2020) Land surface temperature and NDVI, NDBaI and MNDWI were calculated. Correlation analysis was used to indicate relationships between LST with NDVI, NDBaI and MNDWI. The study found that Land Surface Temperature was increased by 50C from 1990 to 2020. Vegetation areas (NDVI) and Water bodies (MNDWI) have strong negative relationship with Land Surface Temperature (R2= 0.99, 0.95) whereas Barren land (NDBaI) has positive relationship with Land Surface Temperature (R2= 0.96). Finally, we recommend the decision makers and environmental analyst to emphasis the importance of vegetation cover and water body to minimize the potential impacts of land surface temperature.


Introduction
Most of human actions are the main cause for constantly declining the vegetation cover of the earth's surface . The decrements of vegetation cover the cause for the increment of land surface temperature. Land surface temperature (LST) defined as skin temperature of the earth's temperature is a key factor for land surface processing studies at both a regional and global scale (Thanh et al., 2018). According to Alemu (2019) stated that land surface temperature is calculated from emitted energy by earth surface or satellite based devices. Relationships of land surface temperatures with Normalized Difference Index, Normalized Barren Index and Modified Normalized Difference Water Index were examined (Qinqin, 2012). Analysis of spatial changeability of Normalized Difference Vegetation Index (NDVI), Land Surface Temperature (LST) the nexus between the factors is essential in natural and environmental studies (Zareie et al., 2016) as well as for spatial decision making and for monitoring of the natural resources. Furthermore, the investigation of Land surface temperature with NDVI, NDBaI and MNDWI of different times can be used to identify land use/land cover changes, which were made because of the forest fires, deforestation, built up expansion, expansion of agricultural land and grazing land restoration (Zhou and Wang, 2011;Mimbrero et al., 2014). Remote sensing indices are employed to quantitatively represent land use/land cover types. For example, the NDVI has been used to confirm the role of green space (Yuan and Bauer, 2007). The Normalized Difference Barren Index (NDBaI) and Modified Normalized Difference Water Index (MNDWI) have been employed to represent degraded land and water areas quantitatively (Gao, 1996;Zha et al., 2003). Though previous studies have been adopted these indices to model Land Surface Temperature (LST) (Chen et al., 2016), only a few of them have compared the modeling results of different years. The study area was experienced increasing Land Surface Temperature (LST) due to degradation of vegetation cover and blue spaces. Therefore, this study attempted to analysis of Spatiotemporal Land Surface Temperature in Relative to Different Indices (NDVI, NDBaI and MNDWI) using Geospatial Techniques in three selected districts (Gida Kiremu, Limu and Amuru), western Ethiopia.

Materials and Methods
Description of the study area This research conducted in selected woreda was located in East Wollega zone and Horo Guduru Wollega zone. Geographically the study area was lies between 9 0 27′00″ and 10 0 18′00″ N and 36 0 19′30″ and 37 0 10′30″ E. Administratively, Amuru district is located in Horo Guduru Wollega Zone and Gida kiremu and Limu districts were located in East Wollega Zone of Oromia National Regional State in Western Ethiopia ( Figure 1). The altitude of the study area lies between 2496.61m and 713.32m above mean sea level. It covers an area of about 5086.65km 2 . Figure 1: Location map of the study area

Types of data and Descriptions
Three different years of landsat imageries were used for this study. Thermal and multispectral bands of landsat TM of 1990, landsat ETM+ of 2003 and landsat OLI/TIRS of 2020 were downloaded from USGS (http://earthexplorer. usgs.gov/) free of cost. The detail information of these data were presented in (Table 1) (Figure 2).

Vegetation covers (NDVI)
This index is used to estimate the amount of vegetation cover on the earth surface. This NDVI was calculated from multi spectral bands of landsat images of all study period. Band 4 for landsat 5 &7, band 5 for landsat 8 used to measure near infrared bands. As well as band 3 for landsat 5&7, band 4 for landsat 8 was used to measure red bands of the landsat data. The formula of this index presented in (Eq 1) = (Eq 1) The NDVI values ranges between -1.0 and 1.0. The NDVI values for health and dense vegetation always ranges between 0.2 and 0.9 (Bustos and Meza, 2015). Whereas, vegetation such as rock, water and barren lands indicated by the values less than 0.1 (Fu and Burger, 2015) Modified Normalized Difference Water Index (MNDWI) The MNDWI is designated to represent water areas often reveal remarkable differences in thermal characteristics when modeling thermal environment (Xu, 2008;Zhifeng and Jianjun, 2012). It calculated from the measurements of green band (band 2 for landsat 5 & 7, band 3 for landsat 8) and middle infrared (band 5 for landsat 5 & 7, band6 for landsat 8) reflectance the formula (Eq 2). = (Eq 2)

Normalized Difference Barren Index (NDBaI)
The NDBaI is selected to represent barren land areas often indicate significant differences in thermal characteristics when predicting thermal environment (Zhifeng and Jianjun, 2012). It calculated from the middle infrared (band 5 for landsat 5 and 7, band6 for landsat 8) and thermal infrared (band 6 for landsat 5 and 7, band 10 and 11 for landsat 8) reflectance, calculated by formula (Eq 3). = (Eq 3)

Retrieval of Land Surface Temperature (LST)
Land surface temperatures (LST) developed from Landsat ETM+ and OLI/TIRS data, including mono window algorithm (Qin et al., 2001), single channel and split window algorithm (Jimenez and Sobrino, 2003). Split window algorithm was used to retrieve LST from Landsat 8 data that has two bands (Band 10 and Band 11). It uses brightness temperature of the two bands of Thermal Infrared (TIR), mean and difference in land surface emissivity for estimating LST (Cheng et al., 2015).
Step I: Conversion of DN in to Radiance A. Mono Window Algorithm In Mono window algorithm digital numbers are converted to at-sensor radiances sensor, before to calculate brightness temperature. The ETM+ DN values ranges between 0 and 255 (Eq.4). LMINλ= is the spectral radiance that is scaled to QCALMIN in watts/(meter squared*ster* μm) LMAXλ= is the spectral radiance that is scaled to QCALMAX in watts/( meter squared*ster* μm) QCALMIN= is the minimum quantized calibrated pixel value corresponding to LMINλ in DN QCALMAX= is the maximum quantized calibrated pixel value corresponding to LMAXλ in DN= 255 B. Split window algorithm The split window algorithm (SWA) is used for Landsat 8 LST estimation as previously used by various scholars (Aik et al., 2020;Atitar and Sobrino, 2009). The Digital Numbers (DNs) of bands 10 and 11 from the Landsat 8 OLI was first converted to spectral radiance (Eq.5. = ( * ) + (Eq.5) Where; Lλ is Top of Atmosphere (TOA) spectral radiance (Wm −2 sr −1 μm −1 ) ML is Band-specific multiplicative rescaling factor from the metadata (RADIANCE_MULT_BAND x, where x is the band number) AL is Band-specific additive rescaling factor from the metadata (RADIANCE_ADD_BAND_x, where x is the band number) QCal is Quantized and calibrated standard product pixel values (DN) Step II. Conversion to Temperature (ETM + ) The mono-window algorithm is used to calculate LST based on land surface emissivity, atmospheric trans-emissivity, brightness temperature, and average atmospheric temperature (Zhang et al., 2006). TM and ETM+ Band 6 imagery can also be converted from spectral radiance (as described above) to a more physically useful variable (Eq.6). The conversion formula is: = (Eq.6) T = Effective at-satellite temperature in Kelvin K2 = Calibration constant 2 K1 = Calibration constant 1 L = Spectral radiance in watts/(meter squared * ster * μm) The split-window algorithm (SW) was applied to calculate LST for this study. Due to the significant reliability of the SW algorithm, it has a wider applicability. It uses the brightness temperatures of two TIR bands (band 10 and band 11 of Landsat-8) to calculate mean land surface emissivity, and then to estimate LST.
TB10 is brightness temperature of band 10 (Kelvin K); TB11 is brightness temperature of band 11 (Kelvin K); ℇ is mean value of Land Surface Emissivity (LSE) of TIR bands; W is content of water vapors in the atmosphere; ∆ℇ is difference between LSE of bands 10 and 11.e emissivity, difference of land surface emissivity, and then to estimate LST (Eq.7): = ( )
Finally, land surface temperature results from Landsat ETM+ and OLI/TIRS was converted into degree Celsius, by subtracting 273.15. To convert temperature the degree Kelvin ( 0 K) to degree Celsius ( 0 C), (Eq 11) was used. Gradually, LST was increased in all selected year the mean of 23. 70C in 1990, 24.30C in 2003 and 28.70C in 2020. The average temperatures rise from 1990-2020 by 5 0 C this increment is between 30 years. According to Rasul et al. (2012) during the 20 th century, they estimate the temperature will increase further between 1.4 0 C and 5.8 0 C by 2100 year.   Generally, the relations of all parameters were presented in (Table 2). Land surface temperature has direct relationship with Normalized Difference Barren Index (NDBaI) while LST has indirect relationship with NDVI and MNDWI.