1 Land Surface Temperature Measurements from EOS MODIS Data

• Main Author: Zhengming Wan
• Other Authors: The Pennsylvania State University CiteSeerX Archives
• Format: Text
• Language: English
• Published: 2003
• Subjects: Greenland
• Online Access: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.666.372; http://icess.ucsb.edu/modis/wan2003_2.pdf

A cloud-screen scheme with constraints on spatial and temporal variations in LSTs is developed to screen off the cloud-contaminated LSTs in the MODIS LST products. This scheme is applied to the V3 and V4 1km MODIS LST (MOD11A1) products in Lake Tahoe and Namco (a lake in Tibet) in 2002 to show the impact of using the MODIS cloudmask (MOD35) product in different ways: processing lake and river pixels only in clear-sky condition at the 99 % confidence level in V3 and at equal to or larger than the 66% confidence level in V4. The test study unveils that cloud-contaminated LSTs exist even at pixels defined by MOD35 in clear-sky conditions at the 99 % confidence level and that these contaminated LSTs can be removed by the cloud-screen scheme. The cloud-screen scheme is also applied to the V4 5km MODIS LST (MOD11B1) product at test sites in lakes, snow/ice, Amazon rain forest and semi-arid regions, indicating that the MODIS cloudmask needs to be improved. After screening off the cloud-contaminated LSTs, the time sequence of the MODIS LST products clearly shows the freezing and lake-ice melting in Namco, and the snow/ice melting period in Greenland. The MOD11B1 product is validated through comparing the LSTs retrieved by the day/night method to the LSTs in the MOD11A1 product that has been validated with in-situ measurement data in lakes, snow and agricultural fields in field campaigns conducted in 2000-2003. The mean and standard deviation values of the differences between the LSTs retrieved by