How to Use ArcGIS Pro to Calculate Land Surface Temperature (LST) from Landsat Imagery

| |

Land Surface Temperature (LST) is defined by the National Aeronautics and Space Administration (NASA) Earth Observatory as how hot the “surface” of the Earth would feel to the touch in a particular location. 

Many areas of research, including global climate change, hydrological and agricultural processes, and urban land use/land cover, increasingly rely on LST. The advent of remote sensing images has eased the calculation of Land Surface Temperature spatiotemporally; through its thermal radiance as LST is a significant factor in controlling environmental, biological, chemical, and socio-economical processes on the earth.

Consequently, this tutorial utilizes spatial data science and remote sensing techniques (ArcGIS Pro) to access the LST in a region. Normalized Difference Vegetation Index (NDVI), Brightness Temperature (BT), and Land Surface Emissivity (LSE) are factors that will be considered in the calculation process as well as making use of Landsat 7 and Landsat 8 thermal bands. Figure 1 summarizes the procedure that will be adhered to in undertaking this tutorial.

Figure 1: Methodology flow for estimation of LST using ArcGIS Pro.
Figure 1: Methodology flow for estimation of LST using ArcGIS Pro.

Working with Landsat 5 and Landsat 7 Imagery

Landsat 7/Landsat 5 imagery can be downloaded freely from USGS.com. For this tutorial, Landsat 7 imagery will be used. The same steps will hold for the Landsat 5 imagery. After downloading the imagery, line correction will be performed on the imagery. 

The Landsat toolbox can be downloaded from https://drive.google.com/file/d/1FKc-G1vMVtWXi66hh15VKoj-zk3UPWgX/view. After, the toolbox can be imported, and used for scanline correction using the following steps:

  1. Open and create a new project in ArcGIS Pro.
  2. Select insert tab from the menu bar. Click the “add toolbox” icon from the Toolbox drop-down menu. Toggle to the location of the Landsat toolbox to select it. 
Figure 2: Adding a new toolbox in ArcGIS Pro.
Figure 2: Adding a new toolbox in ArcGIS Pro.
  1. Type and search “Fix Landsat 7 Scanline Errors.” Set “input Band with Bad Rows” tab to Band 6 of Landsat 7. 
  2. Set the “output image” to the desired output name and location. Results obtained will be Landsat 7 Band 6 imagery free of scanline errors.

Figure 5 shows the formulas for the calculations undertaken at each stage of the analytical process. Some of the variables as illustrated in figure 4 can only be found in the metadata (MTL) file of the Landsat imagery, which happens to contain all the attribute information of the imagery.

Illustration of the MTL file.
Figure 4: Illustration of the MTL file.

Formulas for Estimated LST from Landsat 5 and Landsat 7

Formulas for estimating land surface temperature (LST)
Figure 5: Formulas for estimating land surface temperature (LST)
table showing formula components for LST formula for Landsat 5/7.
Explanation for each of the formula components.
Explanation for each of the formula components.

Outlined below are steps to follow in undertaking the formulations in Figure 5 in ArcGIS Pro:

  1. Type and search for “raster calculator” from the geoprocessing search bar. 
  2. Use the calculation as shown in Figures 5 and 6 to convert Digital Numbers (DN) values to Top of Atmosphere radiance.
  3. Set output raster to the desired name and location.
  4. Click, “Run.”
Figure 6: Conversion of DN values to TOA radiance in ArcGIS Pro.
Figure 6: Conversion of DN values to TOA radiance in ArcGIS Pro.

In order to convert Radiance to Brightness Temperature, the following steps should be adhered to;

  1. Type and search raster calculator from the search bar of the geoprocessing toolbox.
  2. In the calculation bar of the map algebra (raster calculator), key in the formulas as shown in figures 5 and 7.
  3. Set “output raster” to the desired name and location.
  4. Click, “Run.”
Figure 7: Conversion from radiance to BT in Kelvin in ArcGIS Pro.
Figure 7: Conversion from radiance to BT in Kelvin in ArcGIS Pro.

To convert from Kelvin to Degree Celsius, subtract 273.15 from the results of “BT in Kelvin” as follows, and click, “Run.”

Figure 8: Conversion from Kelvin to degree Celsius in ArcGIS Pro.
Figure 8: Conversion from Kelvin to degree Celsius in ArcGIS Pro.

Working with Landsat 8 Imagery

After downloading the imagery, line correction should be performed on the band 10 (thermal infrared) of the imagery as illustrated above.

In order to convert the Digital Numbers (DN) to Top of Atmosphere (TOA) radiance, the following steps should be taken;

  1. In the search bar of the geoprocessing toolbox, type and search, “raster calculator.”
  2. The following formula (figure 9) should be typed into the calculation bar of the map algebra (raster calculator).
  3. Set “output raster” to the desired name and location.
  4. Click “Run.” 
Figure 9: Conversion from DN values to TOA radiance in ArcGIS Pro.
Figure 9: Conversion from DN values to TOA radiance in ArcGIS Pro.

In order to convert the thermal band from TOA Radiance to Brightness Temperature (BT), the following steps should be followed;

  1. In the search bar of the geoprocessing toolbox, type and search “raster calculator.”
  2. Type the BT formula into the calculation bar of the map algebra (raster calculator).
  3. Set “output raster” to the desired name and location.
  4. Click “Run.”

                                                               BT= 

From the formula in figure 10, “-273.15” is added, to convert the results from Kelvin to Degree Celsius. 

Figure 10: Illustration of brightness temperature (BT)
Figure 10: Illustration of brightness temperature (BT)

Table 1 also summarizes the inferences that will have to be made on the “MTL” extension file (metadata) in undertaking the calculation process. 

VariableDescriptionValue
K1 K2Thermal constants, Band 10774.8853 1321.0789
Lmax LminMaximum and Minimum values ofRadiance, Band 1022.00180 0.10033
Qcalmax QcalminMaximum and Minimum values ofQuantize Calibration, Band 1065535 1
OiCorrection value, Band 100.29
Table 1: Some data values from the metadata (MTL) of the Landsat 8 imagery.

To calculate NDVI, 

  1. Band 4 and Band 5 of Landsat 8 should be imported onto the map canvas by the “Add data” tool after line correction had been performed on them.
  2. Raster calculator should be typed and searched from the search bar of the geoprocessing window.
  3. The formula below should be typed into the calculation bar of the raster calculator.

NDVI= (Band 5 – Band 4)/ (Band 5+ Band 4)

Proportional value vegetation will be calculated from NDVI, using the following formula and steps; 

  1. Launch the raster calculator.
  2. Input the following formula as shown in figure 9 into the calculation bar.
  3. Set “output raster” to the desired name and location.
  4. Click “Run.”
Raster calculator GUI for proportional value vegetation calculation in ArcGIS Pro.
Figure 12: Illustration of proportional value vegetation.

PV=  

The following formula should be inputted into the calculation bar of the raster calculator in order to calculate Land Surface Emissivity (LSE). Click, “Run.”

Formula for calculating Land Surface Emissivity (LSE).
Formula for calculating Land Surface Emissivity (LSE).
Raster calculator with formula for and Surface Emissivity (LSE) in ArcGIS Pro.
Using the raster calculator in ArcGIS Pro to calculate Land Surface Emissivity (LSE).

LST can be calculated from LSE, BT and NDVI. The following steps will be adhered to;

  1. Open raster calculator. 
  2. Set “output raster” to the desired name and location.
  3. The following formula should be inputted into the calculation bar of the raster calculator window.
  4. Click “Run.”
Formula for calculating land surface emission from LSE, BT and NDVI.
Formula for calculating land surface emission from LSE, BT and NDVI.
Figure 14: Illustration of LST calculation.
Figure 14: Illustration of LST calculation.

Clipping of LST to Study Area

To limit the extent of the Landsat imagery to an area of interest, follow the steps below;

  1. Type and search “extract by mask” from the search bar of the geoprocessing toolbox.
  2. Set “input raster” as LST Landsat 7. 
  3. Set “output raster” as the desired name and location.
  4. Click “Run.”
  5. Repeat steps for LST Landsat 8.
LST map of Connecticut, USA (Landsat 7).
Figure 15: LST map of Connecticut, USA (Landsat 7).
Figure 16: LST map of Connecticut, USA (Landsat 8)
Figure 16: LST map of Connecticut, USA (Landsat 8)

Further Resources

For more information on LST inverse with the radiative transfer equation procedure see: Yu, X., Guo, X., & Wu, Z. (2014). Land surface temperature retrieval from Landsat 8 TIRS—Comparison between radiative transfer equation-based method, split window algorithm and single channel method. Remote sensing6(10), 9829-9852. https://doi.org/10.3390/rs6109829

More ArcGIS Pro Tutorials

Related

Share this article


Enter your email to receive the weekly GIS Lounge newsletter: