Hybrid modelling of random forests and kriging with sentinel-2A multispectral imagery to determine urban brightness temperatures with high resolution

ABSTRACT

Herein, a random-forest-based Kriging downscaling method (RFRKriging) was tested to increase the spatial resolution of Landsat 8 thermal-infrared remote sensing (TIRS) brightness temperature (T_b) over the central areas of Guangzhou, China. The 240-m Landsat 8 T_b was upscaled from the resampled 30-m Landsat 8 T_b by using the spatial averaging method. T_b was downscaled based on its relationship to spectral reflectance in visible/near-infrared and shortwave infrared bands, spectral indices, including normalized difference vegetation index (NDVI), normalized difference building index (NDBI), and modified normalized difference water index (MNDWI), digital elevation model (DEM), longitude, latitude, impervious surface, vegetation, and soil fractions. The linear regression Kriging (LRKriging) and NDVI-based thermal sharpening algorithm (TsHARP) regression Kriging (TsHARPKriging) were also implemented. Relative to the reference data from the resampled Landsat 8 T_b, the RFRKriging produced satisfactory downscaling results with respect to the maximum coefficient of determination (R²), the minimum root-mean-square error (RMSE), and the minimum ratio between RMSE and the correspondent Landsat 8 T_b standard deviation (RSD). RFRKriging captured T_b differences over roads, forests, water bodies, and buildings and enhanced the spatial details of downscaled T_b in all urban areas. Furthermore, integrating the multi-type predictor variables from 10-m Sentinel-2A and 30-m Landsat 8 image sets, RFRKriging was further applied to downscale 30-m Landsat 8 T_b to 10 m. Then, it was upscaled to 30 m and assessed with the original 30-m Landsat 8 T_b. RFRKriging was an effective and practical method for increasing the spatial resolution, with a low RSD of 0.254, a low RMSE of 0.525 K and a high R² of 0.935. The RSD of less than 0.5 indicates that RFRKriging could achieve an acceptable downscaling accuracy. Furthermore, RFRKriging enhances the spatial details of a 10-m T_b in complex urban environments while maintaining the intrinsic information and spatial patterns of a downscaled T_b. Due to the unavailability of the 30-m and 10-m T_b truth, this study may only provide a quantitative assessment from some theoretical experiments. Also, a comparison of the RFRKriging downscaled T_b with the resampled Landsat 8 T_b was performed. The results suggest that the proposed RFRKriging downscaling method shows better downscaling performance with more spatial details than the cubic convolution method applied by the National Aeronautics and Space Administration (NASA) Centre.

KEYWORDS:

• Downscaling
• Landsat 8 TIRS brightness temperature
• Random forest
• Ordinary Kriging
• Sentinel-2A

Acknowledgements

This work was supported in part by the Science and Technology Planning Project of Guangdong Province (No. 2018B020207012), the Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (No. GML2019ZD0301), the National Natural Science Foundation of China (Nos. 41901371 and 41801369), the Science and Technology Planning Project of Guangdong Province (No. 2018B020207002), the Guangdong Innovative and Entrepreneurial Research Team Program (No. 2016ZT06D336), and the GDAS’ Project of Science and Technology Development (Nos. 2018GDASCX-0403, 2019GDASYL-0301001, and 2019GDASYL-0501001).

Disclosure statement

The authors declare no conflict of interest.

Additional information

Funding

This work was supported by the Science and Technology Planning Project of Guangdong Province [2018B020207012]; Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) [GML2019ZD0301]; National Natural Science Foundation of China [41901371, 41801369]; Science and Technology Planning Project of Guangdong Province [2018B020207002]; Guangdong Innovative and Entrepreneurial Research Team Program [2016ZT06D336]; GDAS' Project of Science and Technology Development [2018GDASCX-0403, 2019GDASYL-0301001, 2019GDASYL-0501001].