ASTER Surface Emissivity Product
Version 2.9, June 2002

This Validated version of the Surface Emissivity product provides an estimate of the thermal emissivity for each of the five ASTER TIR bands. This version is a Validated version intended to provide a data product that we have found to be computationally correct and whose accuracy has been estimated (see below). This product is suitable for use in publications given that the user understands the accuracy of the product is dependent on the accuracy of the surface leaving radiance TIR product (AST09T), which in turn is dependent on the accuracy of the atmospheric profiles used in the correction procedure.

This product iteratively corrects the calculated land-leaving radiance for the effects of downwelling atmospheric irradiance as it calculates and refines the temperature and emissvities. The results are primarily limited by the accuracy of the land-leaving radiance product AST09T, which in turn is limited by the accuracy of the atmospheric compensation. Under ideal conditions, the inaccuracies from the atmospheric correction are about the same as inaccuracies from two other sources: image measurement errors, and errors attributable to the scaling regression relating scene emissivity to emissivity contrast (min-max difference, or MMD). (Later versions of this product will be improved by using internal checks on the effectiveness of the atmospheric correction, and using input land-leaving radiance product AST09T calculated using atmospheric temperature and water vapor profile information derived from MODIS, a sister instrument to ASTER on the TERRA platform.) 


Table of Contents

1. Where to get detailed information on this product
2. How to report problems
3. Description of this version
    * Product Quality
    * Problems and limitations
4. Version history


1. Where to get detailed information on this product

There are several sources of information that users will find useful:

*"Temperature/Emissivity Separation Algorithm Theoretical Basis Document, Version 2.4." is the title of the theoretical basis document (ATBD) for this product. This document describes the algorithm used to create the product. It is available in pdf format as document atbd-ast-03 at the EOS Project website: http://eospso.gsfc.nasa.gov/eos_homepage/for_scientists/atbd/viewInstrument.php?instrument=24

*Gillespie, A. R., Matsunaga, T., Rokugawa, S., and Hook, S. J., 1998, Temperature and Emissivity Separation from Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Images. IEEE Transactions on Geoscience and Remote Sensing, 36, 1113-1126.

* ASTER Higher-Level Product User Guide. This document describes each of the ASTER Higher Level products. It is available in pdf format at the ASTER site: http://asterweb.jpl.nasa.gov/documents/documents.htm

* US ASTER Website. This site describes the ASTER instrument including provision of the relative spectral response for each band, how to obtain data, the various data products that are available, the activities of the US ASTER Science Team, sample images and other ASTER information The address for this site is: http://asterweb.jpl.nasa.gov/

* Level 2 product information. Because the TIR Temperature and Emissivity Product is calculated from the Surface Leaving Radiance product, understanding the AST09 product is useful. 

* "An Atmospheric Correction Method for ASTER Thermal Radiometry Over Land" is the title of the theoretical basis document (ATBD) for this AST09. This document describes the algorithm used to create the product. It is available in pdf format as document atbd-ast-05 at the EOS Project website: http://eospso.gsfc.nasa.gov/eos_homepage/for_scientists/atbd/viewInstrument.php?instrument=24

2. How to report problems

To report a problem with the product, please send mail to edc@eos.nasa.gov or call EDC DAAC User Services at (605) 594-6116.


3. Description of this version

ASTER Surface Emissivity AST05 Version 2.9

This is the third release of this product and it should be considered a Validated version.

The five Emissivity values are dimensionless proportionality factors. The scaling factor is 0.001. 

In converting image Data Numbers (DN) to Emissivity there are no offsets, and the scaled values are obtained by multiplying the image DN by the appropriate scaling factor (value=DN*scaling factor). 


Changes

Validated Release


Product Quality & Limitations

Extensive validation of the TIR Level 1B (L1B) radiance at sensor product has been conducted by several ASTER groups over the last year and all comparisons indicate performance within ±1% or better. 

The Level 2 surface-leaving radiance and irradiance products are dependent on the quality not only of the at-sensor radiance but the quality of the atmospheric profiles and the topographic data based used. Validation experiments suggest agreement with field data at the 1% level, but there may occasionally be significant disagreement between the atmospheric profiles obtained from the NOAA assimilation model (used in creating the land-leaving radiance) and in-situ measurements, resulting in a 3-6% disagreement between the Product and field measurements. This appears to be the largest factor limiting reliability of the AST03 Temperature and Emissivity Products. A detailed discussion of errors in the Products is available in the ATBD referenced above. 

Under ideal conditions, the Temperature product is accurate within 1.5K and precise within 1.5K at about 300K: that is, temperatures for a large homogeneous scene would be within 1.5K of their correct values, and the value would be repeatable within 1.5K for measurements at different dates. Emissivities are also accurate within 0.015 emissvity units, and precise within 0.015 units. If uncertainty is due to correction for downwelling sky irradiance, the reported temperature may be more accurate than the emissivities, especially for the lower values of emissivities in the spectrum. This is because the emissivities are affected to a greater extent by the downwelling irradiance if the have low vales, whereas temperatures are calculated from the least-affected, highest of the five emissivities. 

The Temperature and Emissivity Products are only valid for cloud-free pixels. The algorithms used do not correct or account for the presence of water or ice clouds and the products are of uncertain value when such clouds are present. Furthermore, in the immediate vicinity of clouds, temperatures and emissivities may be increased because of radiative exchange between land and cloud. A cloud mask is included in the quality assurance "QA plane" portion of the product, allowing the user to avoid cloudy pixels. This cloud mask product is currently not validated. If clouds are of a concern the user should look at the corresponding Level 1B VNIR/SWIR/TIR images for guidance. 

This data product has been moved from a beta version to a provisional version based on evidence that the product is computationally correct and the algorithm functions as it was intended to. This assessment was made using a limited number of field measurements in which the Temperature and Emissivity Products were compared with independent field-based results. Specifically, validation is from experiments conducted over 1) "water" sites, with low spectral contrast (Lake Tahoe; Salton Sea; Pacific Ocean) and 2) "rock" sites with higher spectral contrast (Mauna Loa, 4 km amsl; Mauna Loa - Mauna Kea saddle, 2 km amsl; and Punaluu, at sea level). By testing similar sites viewed through different amounts of atmosphere, error and uncertainity can be parsed between 1) instrumental and algorithm effects, and 2) atmospheric effects. 

Validation of the ASTER Temperature/Emissivity Separation algorithm on a global basis is not feasible as the necessary measurements of surface temperature and surface spectral emissivity and atmospheric properties at the time of ASTER data acquisition are not generally available. A major source of systematic error is natural scatter about the empirical regression line used to scale the emissivity data. Therefore, algorithm performance in any given area with a given set of lithologies needs to be validated independently for best results. 

4. Version History

Version 2.9 June 2002

Version 2.8 April 2002

Version 2.6 October 2001

Version 2.5 April 2001 Initial Release