Modeling for Land Surface Temperature Change in Taiwan

Abstract

Land surface temperature (LST) is an important factor in surface energy balance and global climatology studies. LST characteristics such as elevation, land cover (LC) and vegetation can significantly affect LST. This study aimed to investigate the annual seasonal patterns, decadal trends, their relationships, and dynamic acceleration patterns in LST and NDVI changes through diverse LC types and to examine the influence of LC variation and elevation on LST in a decadal trend in Taiwan by using weighted sum contrasts linear regression. LST, NDVI and LC data were retrieved from the MODIS Land Product subset tool (ORNL DAAC, 2018) and elevation data were downloaded from the USGS Earth Explorer website. The natural cubic spline method and multivariate linear regression were used to model annual seasonal patterns, decadal trends, and acceleration rates for each LU type. Acceleration patterns were analyzed using the correlations between derived trends and acceleration rates. Finally, weighted sum contrasts linear regression was applied to evaluate the influence of LC alteration and elevation on the decadal change of LST. Results showed that BB land had a significant increase in LST (0.862°C per decade). The average increase in daytime LST and NDVI were 0.141°C and 0.019 unit per decade, respectively. Moreover, Taiwan had a significant mean increase in decadal trends of NDVI but not LST. LST and NDVI had a negative relationship and were stronger in green coverage areas compared to built-up areas. At a sub-region level, a higher rise in LST and tremendous vegetation loss were found at high altitudes in the Central Mountain Range, which is located in the southern regions. At the pixel level, urbanization and agricultural expansion had caused increasing LST, while afforestation had contributed to cooling. The study also indicated that the diverse pattern of LC variation has a significant influence on daytime LST, but not on nighttime LST trends. There was an affected of daytime and nighttime LSTs at an altitude above 600 m. However, the growth rates of greenness have not yet decreased, policymakers and practitioners have to consider and plan a balanced view of green cities to retain pace with the accelerating surface temperature conditions.