Nowadays, the topic of climate change is increasingly becoming a topic of discussion, and it should be the case. Economic development and population growth, both directly and indirectly, unknowingly have a significant impact on this global problem. Under current conditions, the IPCC estimates that the temperature increase that will occur between 2030 and 2052 can reach 1,5° C. The use of remote sensing methods can be used to estimate the temperature in an area, one of the most basic is the analysis of calculating the surface temperature of the ground: Land Surface Temperature (LST) Analysis. LST is the average temperature of an element from the earth’s surface which is calculated from the radiation (radiance) and is influenced by the wavelength it captures. The most sensitive wavelength to surface temperature is through a thermal infrared sensor that can be processed from several satellite images, one of which is using Landsat 8 imagery.
Starting from the introduction regarding surface temperature above, the question arises about how the condition of surface temperature is in Kabupaten Bekasi, thich is one of the city that has the stereotype as a hot and congested area. Not to mention that it is the ‘neck’ of the capital which makes it a strategic overseas location, has one of the largest industrial areas in Indonesia and its title as a ‘district’ which oversees villages, if referring to the Spatial Planning Law, have agricultural main activities. Therefore, I conducted an LST analysis in Bekasi Regency using data from 3 different years, namely 2016, 2020 and 2021 to see the difference in temperature in these three years. There are many ways to calculate LST, what is being done here is to use spatial calculations with R and use band 10 imagery from Landsat 8 as one of the bands with the Thermal Infrared Sensor (TIRS). Here are some steps taken.
Download the required image via the USGS website. The data used is Landsat 8 image data for the Kabupaten Bekasi area in 2016 (acquisition time: 13–05–2016), 2020 (acquisition time: 24–05–2020), and 2021 (acquisition time: 11–05–2021).
After obtaining the required data, the next step is to open the metadata file from the data. For example, the 2016 image data file is named LC08_L1TP_122064_20160513_20200907_02_T1_MTL.txt. Just notice the 3 letters before the file extension, it usually says MTL. Then, take the values RADIANCE_MULT_BAND_10 = 3.3420E-04 and RADIANCE_ADD_BAND_10 = 0.10000, and K1_CONSTANT_BAND_10 = 774.8853 and K2_CONSTANT_BAND_10 = 1321.0789 for further analysis.
I am are using RStudio as an IDE. The first step is to import the library needed for this simple analysis: raster. Before that, first make sure the data directory is correct, then input the values obtained from the metadata as in the previous stage.
# import library
library(raster)
# values from metafile
RADIANCE_MULT_BAND_10 <- 3.3420E-04
RADIANCE_ADD_BAND_10 <- 0.10000
K1_CONSTANT_BAND_10 <- 774.8853
K2_CONSTANT_BAND_10 <- 1321.0789
# load raster file
band_10 <- raster("D:/data/Landsat 8 (2016 lvl 1)/LC08_L1TP_122064_20160513_20200907_02_T1_B10.TIF")Convert the Digital Numbers (DN) to Top of Atmosphere (TOA) Reflectance. The trick is to use the calc command. The formula is as follows.
# CONVERT DIGITAL NUMBERS (DN) TO TOP OF ATMOSPHERE (TOA) REFLECTANCE
# calculate TOA from DN
toa_band10 <- calc(band_10, fun=function(x){RADIANCE_MULT_BAND_10 * x + RADIANCE_ADD_BAND_10})Convert the Radiance value to Brightness Temperature (BT) in kelvin units. The formula is as follows.
# CONVERT TO AT-SATELLITE BRIGHTNESS TEMPERATURE
# calculate LST to Kelvin
temp10_kelvin <- calc(toa_band10, fun=function(x){K2_CONSTANT_BAND_10/log(K1_CONSTANT_BAND_10/x+1)})Then, convert the kelvin value to degrees Celsius using a simple formula below.
# CONVERT TO AT-SATELLITE BRIGHTNESS TEMPERATURE
# calculate LST to Kelvin
temp10_kelvin <- calc(toa_band10, fun=function(x){K2_CONSTANT_BAND_10/log(K1_CONSTANT_BAND_10/x+1)})Then, convert the kelvin value to degrees Celsius using a simple formula below.
# calculate LST to Celcius
temp10_celcius <- calc(temp10_kelvin, fun=function(x){x-273.15})
# plot
plot(temp10_celcius)
Result — Plot 1
After all the steps are done, the next step is to export the analysis result file in raster format (.tif) with the writeRaster command. Do the same for the three years you want to analyze: year 2016, 2020 and 2021.
# export raster writeRaster(temp10_celcius, "temp10_c_2016_lvl1.tif")The above stage is the main stage in calculating LST using simple spatial analysis with R, for the next visualization stage QGIS will be used which is one of the open source GIS tools. The following are the results after layouting using the Carto Positron basemap and shapefile data for the Bekasi Regency Administrative Boundaries — PPBW BIG.
Final Results of Kab Bekasi LST Analysis in 2016, 2020 and 2021 after being processed using RStudio and QGIS
From the comparison of the results of the LST analysis in Kabupaten Bekasi in 2016, 2020 and 2021, we can conclude:
The three image data were acquired in May, where the transition season (pancaroba) usually occurs in the Bekasi Regency area.
The temperature range from 2016 is between 16°C — 30 °C.
4 years later, in 2020, the temperature has increased by around 5–6 °C from a range of 22 °C — 34 °C.
Meanwhile, at a not too distant time interval (1 year later), the temperature rises in the range of 1–2 °C. The temperature range in 2021 is between 23°C — 36°C.
From the three maps above, it can be seen that the concentrations of areas with the highest surface temperatures tend to be in Kecamatan Tambun Selatan, Cikarang Barat, Cikarang Utara and Cikarang Selatan. This is directly and indirectly related to the existence of an industrial area there. This also affects the surface temperature in the surrounding area which tends to get hot (yellow and orange).
In 2016, although not as hot as the following years, the concentration of hot temperatures tends to be more evenly distributed in the center to the south, while in 2020 and 2021 the concentration of the hottest temperatures tends to be in the 4 districts mentioned above.
The temperature range in 2020 and 2021 tends to be lower than in 2016, this may be related to the implementation of reduced mobility when the Covid-19 pandemic first broke out.
The area with the coldest surface temperature is actually in the north which is a coastal area and is directly adjacent to the North Java Sea. This is related to land use there, which is mostly rice fields.
This simple analysis is far from perfect, because the analysis is still very basic and does not consider other variables, such as topography, vegetation density and other advanced analysis. Therefore, for more optimal results a more comprehensive and in-depth analysis is needed.
References:
Norman, J. M., and Becker, F., 1995. Terminology in thermal infrared remote sensing of natural surfaces. Remote Sensing Reviews, 12, 159–173.
UU №26 Tahun 2007 tentang Penataan Ruang.
https://www.gis-blog.com/calculation-of-land-surface-temperature-lst-from-landsat-8-using-r/
Cover Photo: Markus Spiske on Unsplash (https://unsplash.com/@markusspiske)
