Spatiotemporal patterns of land surface temperature of Antarctica from MODIS monthly LST (MYD11C3) data

Abstract

Antarctica is the focus of attention because of increasing ice melt in response to climate warming. The changes are studied using the Moderate Resolution Imaging Spectroradiometer land surface temperature dataset. A decorrelation stretch of multi-temporal monthly land surface temperature (LST) imagery for 2009, 2010 and 2011 indicated that only 11.3 percent of the variance is independent of elevation, distance from coastline, and latitude and longitude in Antarctica. The reconstructed land surface temperature imagery presents standardised values indicating the amount it deviates from elevation, distance from coastline, latitude and longitude predicted. Cluster analysis allowed the identification of residual spatial and temporal patterns termed the temperature oscillation phenomenon that comprises a significant increase of LST in the inland regions during the summer period.

Keywords::

• Antarctic
• MODIS
• land surface temperature
• climatic change
• terrain pattern recognition
• decorrelation stretch

Acknowledgements

The author is grateful for and this paper benefited significantly from the comments, corrections and suggestions of Prof. Graeme Wright, and the two anonymous reviewers.