Statistical Model of Land Surface Temperature in Sumatra Island, Indonesia: 2000-2018

Abstract

Global warming could be assessed by estimating the Land Surface Temperature (LST) change in a regional and global area. The aims of this study were (1) to determine seasonal patterns and trends in time-series data LST in Sumatra Island and (2) in Sumatra Island, Indonesia, to study the relationships between LST, elevation, land cover (LC), and the Normalized Difference Vegetation Index (NDVI). The data were obtained from NASA Moderate Resolution Imaging Spectroradiometer (MODIS) website and focused on during 2000-2018. The cubic splines had been applied to investigate the seasonal patterns and time series trends of LST change and simple linear regression had been applied to analyze the relationship between elevation, LC, and NDVI with LST change in Sumatra Island. In 2018, approximately 44.6% of Sumatera area was evergreen broadleaf forest where the urban area only covered 0.51% of the whole area. Cubic spline result identifying the trend indicated that 36.2% of Sumatera experiencing decreasing of LST change and other 27.7% increasing of LST change. Furthermore, linear regression identified the factors which concluded the lowest LST change occurred in elevation 60-109 meter above sea level (MASL) evergreen broadleaf forest with NDVI less than 0.75 and the highest change occurred in 320+ MASL Savannas with NDVI 0.8-1.0. Based on the 3 determinants, LST decreased by 0.122 °C on an overall mean per decade. The findings can be applied to monitor environmental changes, particularly the state of the regional climate change relating to planting crops, agricultural expansion and deforestation. However, the other factors need to be taken into account in the model.