Computing Spatio-temporal variations in land surface temperature: A case study of Tehsil Murree, Pakistan
Keywords:
Land surface temperature; LST; spatio-temporal variation; spatial dependence; geographically weighted regression; GWR; spatial heterogeneity; mountainous terrain; Pakistan;Abstract
Background: Land Surface Temperature (LST) is a significant factor for surface processes. Mountain is critical for global energy and mass balance. Deviations of topographic properties of the land surface are causative factors of LST variations.
Objectives: In this study, the spatio-temporal variations in LST are investigated according to changes in elevation, slope, vegetation (Normalized Difference Vegetation Index - NDVI), rainfall, and snowfall (Normalized Difference Snow Index – NDSI). Tehsil Murree, Punjab was chosen as a case study.
Methods: Digital Elevation Model (DEM) and Images of Landsat 8, Sentinel-2A, and MODIS for each month are used to calculate mean seasonal LST for 2017. The slope and elevation of the study area are extracted from DEM while MODIS is used to compute LST. Sentinel-2 images are used to calculate NDVI, and rainfall is collected from PERSIANN-CCS moreover Landsat-8 is used to calculate NDSI. All independent parameters are correlated with dependent-parameter by using linear regression and geographically weighted regression (GWR).
Results: Both models reveal that elevation is the most significantly controlling the distribution of LST. Other variables also influence LST; e.g., rainfall and vegetation (NDVI - third significant influencer) while using linear regression. The vegetation (NDVI) and rainfall (rainfall - last in rank) affect the LST when evaluated via the GWR model. Slope and NDSI in linear regression and GWR also affect the LST respectively. In general, the minimum and maximum temperatures in summer, monsoon, and winter seasons remained 24ºC and 36ºC (summer), 20ºC and 33ºC (monsoon), and 14ºC and 23ºC (winter), respectively. The mean temperatures remained 30ºC, 26.5ºC, and 18.5ºC in summer, monsoon, and winter, respectively.
Conclusions: This study might be helpful to understand the most significant factors affecting LST and to estimate the water stress, flood, and snowline as well to understand the recent trends of global warming in correlation with the seasonal profile in hilly areas.
