Validation of Satellite-Derived Land Surface Temperature Products – Methods and Good Practice

Monday, 15 December 2014: 5:00 PM
Pierre C Guillevic1, Glynn C Hulley2, Simon J Hook1, Jim Biard3 and Darren Ghent4, (1)Jet Propulsion Laboratory, Pasadena, CA, United States, (2)NASA Jet Propulsion Laboratory, Pasadena, CA, United States, (3)Cooperative Institute for Climate and Satellite North Carolina State, Asheville, NC, United States, (4)University of Leicester, Leicester, United Kingdom
Land Surface Temperature (LST) is a key variable for surface water and energy budget calculations that can be obtained globally and operationally from satellite observations. LST is used for many applications, including weather forecasting, short-term climate prediction, extreme weather monitoring, and irrigation and water resource management. In order to maximize the usefulness of LST for research and studies it is necessary to know the uncertainty in the LST measurement. Multiple validation methods and activities are necessary to assess LST compliance with the quality specifications of operational users.

This work presents four different validation methods that have been widely used to determine the uncertainties in LST products derived from satellite measurements. 1) The temperature based validation method involves comparisons with ground-based measurements of LST. The method is strongly limited by the number and quality of available field stations. 2) Scene-based comparisons involve comparing a new satellite LST product with a heritage LST product. This method is not an absolute validation and satellite LST inter-comparisons alone do not provide an independent validation measurement. 3) The radiance-based validation method does not require ground-based measurements and is usually used for large scale validation effort or for LST products with coarser spatial resolution (> 1km). 4) Time series comparisons are used to detect problems that can occur during the instrument’s life, e.g. calibration drift, or unrealistic outliers due to cloud coverage. This study enumerates the sources of errors associated with each method. The four different approaches are complementary and provide different levels of information about the quality of the retrieved LST. The challenges in retrieving the LST from satellite measurements are discussed using results obtained for MODIS and VIIRS.

This work contributes to the objective of the Land Product Validation (LPV) sub-group of the CEOS Working Group on Calibration and Validation (WGCV) that aims to address the challenges associated with the validation of global land products. It is also part of the EarthTemp initiative which main goal is to develop more integrated, collaborative approaches to observing and understanding Earth’s surface temperatures.