Comparing and Downscaling of Satellite Land Surface Temperature Data over Urban and Suburban Environments.
Abstract
A large amount of the world's population lives in urban areas that are exposed to the threats of climate change on a regional and global level. These regions normally have a distinct different Land Surface Temperature (LST) variations with their surrounding non-urban areas which may cause differences in the air temperature and cause urban heat islands. Therefore, LST in the cities can have a strong effect on human health and way of life. The objective of this study is to develop a high spatial and temporal resolution LST data using a combination of Landsat 8, infrared-based satellite sensors such as the Moderate Resolution Imaging Spectroradiometer (MODIS), and the Geostationary Operational Environmental Satellite-R Series (GOES-16). Landsat 8 Thermal Infrared Sensor (TIRS) provides higher spatial resolution (30 m) estimates of skin temperature every 16 days. MODIS makes daytime and nighttime observations of the Earth's thermal emission at a coarser spatial resolution (1000 m), while GOES-16, which has lower spatial resolution (2 km), measures the skin temperature at a much higher temporal resolution (five minutes). Methods used to blend the products of the three satellites include aggregating (up-scaling) the high-resolution data to a coarser one to examine the systematic differences between them. Seasonal data of Landsat 8 TIRS, MODIS and GOES-16 over New York City and Upstate New York were inter-compared. Minimum and maximum values of LST as well as diurnal temperature amplitude, derived from Landsat 8 TIRS and aggregated at a coarse resolution (1000 m), were compared to MODIS-derived LST. The maximum values of Landsat 8 were found to have higher correlation with MODIS measurements while the mean values of Landsat 8 were found to have higher correlation with GOES-R. For each of these, the agreement was within 2 K. Furthermore, a statistical approach was applied to downscale GOES-16 and MODIS LST products to Landsat 8 spatial resolution every five minutes. The downscaled estimates showed a reasonable agreement (about 1.0 K) when they were validated against independent Landsat images. High resolution LST in urban as well as in suburban regions will improve prediction of heat indices and studying the effect of urban heat islands which are crucial for sustainable and resilient urban environment.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFMGC41G1546B
- Keywords:
-
- 0345 Pollution: urban and regional;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 3322 Land/atmosphere interactions;
- ATMOSPHERIC PROCESSESDE: 0414 Biogeochemical cycles;
- processes;
- and modeling;
- BIOGEOSCIENCESDE: 0493 Urban systems;
- BIOGEOSCIENCES