Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Data and Band Combinations

W.L. Stefanov

ASTER Instrument and Data Format

The ASTER instrument, flying on board the Eearth-orbiting Terra satellite, acquires surficial data in the visible to mid-infrared (or thermal) wavelength regions of the electromagnetic spectrum at a variety of spatial resolutions. Table 1 provides the sensor bandpasses and specific pixel resolutions. ASTER was built by the Earth Remote Sensing Data Analysis Center (ERSDAC) of Japan and is housed on the Terra satellite by a cooperative agreement with NASA's Earth Observing System (EOS) program. It has been in operation since 2000 and has a nominal mission lifetime of six years. Unlike sensors installed on the familiar Landsat satellites that continually collect and transmit data, ASTER is request-driven and therefore a targeted instrument (Abrams, 2000). This means that spatial and temporal coverage over any given area on the Earth’s surface is generally not continuous.

Band	Wavelength (microns)	Spatial Resolution (m)
	Visible to Near-Infrared Bands
1	0.52 – 0.60	15
2	0.63 – 0.69	15
3	0.76 – 0.86	15
	Shortwave Infrared Bands
4	1.60 – 1.70	30
5	2.145 – 2.185	30
6	2.185 – 2.225	30
7	2.235 – 2.285	30
8	2.295 – 2.365	30
9	2.360 – 2.430	30
	Mid-infrared (Thermal) Bands
10	8.125 – 8.475	90
11	8.475 – 8.825	90
12	8.925 – 9.275	90
13	10.25 – 10.95	90
14	10.95 – 11.65	90

Table 1. Wavelength ranges and spatial resolutions of ASTER bands (Abrams, 2000).

There are two groups of ASTER data included in our database. The first group is designated “L1B” and ncludes geometrically and radiometrically corrected at-sensor values. The L1B data contain all fifteen bands of data, and have not been atmospherically corrected. The second group is the “Level 2” data, all of which have been atmospherically corrected. The use of atmospherically corrected data is recommended for detailed spectral analysis and temporal change investigations. Atmospheric correction has been applied to these data using a lookup-table based algorithm developed by the ASTER science team (Jet Propulsion Laboratory, 2001; http://asterweb.jpl.nasa.gov/documents/ASTERHigherLevelUserGuideVer2May01.pdf). Level 2 datasets available in our database include visible through shortwave infrared reflectance (“AST_07”), midinfrared emissivity (“AST_05”) and surface kinetic temperature (“AST_08”).

Available Band Combinations

The following section details the various band combinations available through this site. Some guidelines as to what the various color combinations indicate are also provided, but it should be noted that these are generalizations and may not be accurate for any given pixel. Available data products include non-stretched (i.e. no color enhancement has been applied) and stretched versions. Stretched versions have had a 2% linear stretch applied, meaning the lowest and highest 1% of the image histogram (the distribution “tails”) have been set to 0 and 255 respectively, with the remainder of the histogram recalculated accordingly.

Data product descriptions listed below are relevant for both L1B and Level 2 datasets. For most mapping and visualization purposes, L1B data products are sufficient for use. An exception to this general rule of thumb is when significant atmospheric effects are visible in a scene (i.e. clouds, haze). In this case, it is preferable to use Level 2 data as input into the processing algorithms.

Data products are output in Geotiff format. This format is suitable for direct input into image-processing and GIS environments for further analysis and fusion with other data.

Visible to Near-Infrared

This data product is comprised of bands 3, 2, 1 as Red-Green-Blue (RGB) at 15 m/pixel spatial resolution. This band combination highlights actively photosynthesizing vegetation in red (near-infrared band), with undisturbed bedrock and soils primarily expressed as browns, greens, and greys. Built materials and regions typically exhibit blue-green, reddish - purple, and white colors.

Normalized Difference Vegetation Index (NDVI)

This data product is the result of ((band 3 – band 2)/(band 3 + band 2)) calculated for each image pixel in a given scene. The resulting greyscale image provides a relative abundance map of actively photosynthesizing vegetation at 15 m/pixel. Bright pixels correspond to higher relative vegetation abundance, while dark pixels correspond to lower vegetation abundance.

Shortwave Infrared

These images comprise bands 8, 6, and 4 as RGB at 30 m/pixel. As multiband data in this wavelength range tends to be highly correlated, a decorrelation stretch has been applied to obtain maximum separation in data values between bands. This processing step is applied for both the non-stretched and stretched data products. These bands have been selected primarily to highlight spectral features diagnostic for iron oxides, illite, and kaolinite (bands 8 and 6); and carbonates (band 4). This data product is designed for rapid reconnaissance based on these general mineral types; multispectral analysis using all six calibrated shortwave bands is recommended for detailed mineralogical investigations.

Mid-infrared (Thermal)

Bands 13, 12, 10 as RGB at 90 m/pixel comprise this data product. As multiband data in this wavelength range tends to be highly correlated, a decorrelation stretch has been applied to obtain maximum separation in data values between bands. This processing step is applied for both the non-stretched and stretched data products. These bands have been selected primarily to highlight spectral features diagnostic for silicates (band 13), iron- and magnesium-bearing minerals and lithologic types (band 10), and carbonates (all three bands). Using these band combinations, quartzites are bright red; basaltic rocks are blue; granitoids are purple-violet; and carbonates tend to be green to yellow-green. For spectral analysis and comparison, the mid-infrared emissivity (“AST_05”) data set is most useful.

Surface Kinetic Temperature

This dataset is derived from atmospherically corrected midinfrared data using an emissivity-temperature separation algorithm developed by the ASTER science team (Jet Propulsion Laboratory, 2001; http://asterweb.jpl.nasa.gov/documents/ASTERHigherLevelUserGuideVer2May01.pdf). Data values are in degrees C, and images have a 90 m/pixel spatial resolution. Greyscale images show high temperatures as bright pixels and low temperatures as dark pixels. This data set is most appropriate for temporal investigations or those studies that involve absolute surface temperature.

Reference

Abrams, M. (2000) The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER): Data Products for the High Spatial Resolution Imager on NASA’s Terra Platform. International Journal of Remote Sensing 21: 847-859.

Validation, Accuracy, and Additional Information

The ASTER Science Team maintains technical documents, user guides, and data product validation information at http://asterweb.jpl.nasa.gov.

Searches for entire individual ASTER scenes using a variety of criteria (including the granule IDs generated by our database) can be performed using the Earth Observing System Data Gateway (http://edcimswww.cr.usgs.gov/pub/imswelcome/), search tools available at ERSDAC (http://imsweb.aster.ersdac.or.jp/ims/html/MainMenu/MainMenu.html), or the USGS Global Visualization Viewer (GLOVIS; http://glovis.usgs.gov/). There is a fee for ordering ASTER data unless you are an approved NASA-affiliated researcher.