Sorry, you need to enable JavaScript to visit this website.

Frequently Asked Questions - Precision Terrain Corrected (AST_L1T)

Why are there negative latitudes in the AST_L1T Geotiffs when the projection is listed as UTM?

A. The standard projection used is basically Transverse Mercator (TM) with all parameters set equal to those of UTM except the false northing for southern hemisphere scenes.  The false northing is set to zero for both northern and southern hemisphere scenes in order to support easy overlay and mosaicing of scenes spanning the equator.  All AST_L1Ts are mapped onto a pixel-based grid such that all pixel locations (in UTM coordinates) are divisible by the pixel size.  This ensures that pixels in common regions for multiple scenes will have matching locations, thus removing the necessity of resampling. Also most tools handle UTM with this variation better than TM, which is UTM in all ways but false northing of latitudes.

Why do my polar scenes look distorted?

The UTM projection, the chosen projection for AST_L1Ts, is distorted near the poles.

Why is there a red stripe on the right side and a green stripe on the left side of many of the Full Resolution images?

The red and green stripes occur on the right and left sides of AST_L1T scenes containing SWIR bands.  The image RGB bands consist of SWIR band 4 (red), VNIR band 3N (green), and VNIR band 2 (blue).  The AST_L1A data for these bands have different ground “foot prints”, most likely due to slightly differing telescope ground views.  The bands used in the color images are not trimmed to the common ground coverage because data in the trimmed regions may be of interest to some users.

What bands are used for the Full Resolution images?

For the Thermal image, the following bands are used:

Red: band 14
Green: band 12
Blue: band 10

For the Visual image, the bands used depend on the bands available in the AST_L1T.  If SWIR and
VNIR are available:

Red: SWIR band 4 (resampled to 15 meter resolution)
Green: VNIR band 3N
Blue: VNIR band 2

If SWIR is not available:

Red: band 2
Green: band 3N
Blue: band 1

 

What does the color represent in a Thermal Full Resolution image?

The VNIR, TIR, and QA Browse files may not be present for all AST_L1T scenes since they are dependent on the VNIR and TIR telescopes being turned on and whether a geolocation verification is attempted. The individual browse files end in _BR_#.<browse type>.jpg where the # may be  2, 3, or 4, and the browse type may be VNIR, TIR or QA. The numbers represent the order the browse files were put into an HDF file for Ingest. The Browse files are put into the HDF file with VNIR first, TIR second and QA third. For granules that have all 3 Browse files the file names would end: _BR_2.VNIR.jpg _BR_3.TIR.jpg and _BR_4.QA.jpg. If the VNIR telescope is not on, the VNIR browse file is not created and the TIR file will end in _BR_2.TIR.jpg. If the VNIR telescope is on and the TIR telescope is off and the QA browse is created, then the QA browse file name will end with _BR_3.QA.jpg. 

Why are the corners clipped for some AST_L1T VNIR only scenes?

In such scenes, the information used for mapping from AST_L1A to AST_L1B space contains errors which project the left and right boundary data outside the scene frame (upper left, upper right, lower left, and lower right bounds).  The AST_L1T process uses the AST_L1A to AST_L1B mapping parameters for systematic corrections such as band offsets, so such errors are inherited by the AST_L1T from the AST_L1B. This clipping is more obvious in the AST_L1T since the scene is rotated to UTM grid north up, thereby rotating the clipping across the corners.

Why isn't there an AST_L1T for every AST_L1A granule?

AST_L1Ts are not produced if any of the available AST_L1A bands fail during resampling.  When this occurs, it’s typically due to large errors in the AST_L1A geolocation arrays.  There are also scenes containing errors that cause failures in the AST_L1B processing.  Since AST_L1T processing uses the AST_L1B processing algorithms, AST_L1B failures will also be AST_L1T failures. Note: A band is not available if the telescope is turned off.

Why aren't all scenes precision corrected?

Precision correction is based on the ability to match reference ground chips (small regional image data) to an equivalent region in the AST_L1T scene.  There are multiple situations where such a match may not be possible:

  • The region is obscured by clouds.
  • The scene is dark (possibly a night scene).
  • The sun angle causes enough terrain shadowing to make the region look significantly different from the reference chip.
  • Only SWIR and VNIR bands may be used for chip matching because of the relative similarity of the radiometric signatures.  So TIR-only scenes cannot be precision corrected.

The precision correction algorithm requires that a minimum number of ground control point chips be matched in order to generate a reasonably accurate precision fit model for the scene.  This minimum is currently set to 20.  So if at least 20 points are not matched, the scene will “fall back” to a systematic correction (basically a north-up AST_L1B).

Why aren't all scenes terrain corrected?

The only areas that don’t have DEM data available are over ocean water and small islands. These areas will not have terrain correction performed.

Where can I find the path/row the scene is at?

WRS path/row information isn’t usually associated with ASTER scenes, at least not to the extent it is for other datasets such as Landsat scenes.  Since precision processing makes use of Landsat GLS scene information from specific WRS-2 paths and rows, it’s reasonable to make that association for AST_L1T scenes where geolocation verification has been performed.  The path/row information is available in the header of the associated QA product text file for these scenes.

Why do the AST_L1T products start at Version 003?

The initial AST_L1T is generated from AST_L1A.003 input, so it is considered a version three product.   Reprocessing of the AST_L1T product resulting from changes to the AST_L1A or AST_L1T algorithms will result in incrementing the version.

The AST_L1T product incorporates the radiometric and geometric corrections integrated in the AST_L1B at-sensor radiance product, in addition to being orthorectified. The AST_L1B product will continue to be offered during the initial release of the AST_L1T product, however subsequent analysis and potential changes to the production of higher level ASTER products may result in the eventual discontinuation of the AST_L1B product.

How can I tell if the product is terrain or precision corrected?

The AST_L1T metadata contains the field ‘CorrectionAchieved’ that is populated with the level of correction (e.g. Terrain+Precision, Terrain+Systematic, Systematic, or Precision) obtained for each scene / granule.

What does terrain correction do?

Terrain correction removes (or lessens) the geometric errors associated with observing a ground location from an off-nadir angle.  All locations within an AST_L1T are assumed to be viewed from directly above the geoid (nadir), with no perspective (ortho-rectified).  But since practically all of the scene locations are viewed from off-nadir, the elevation of a point makes its location appear farther from satellite nadir than it should.

What does the precision correction do?

Precision correction removes (or lessens) the geometric errors due to imprecise knowledge of the satellite ephemeris (location and velocity), attitude (yaw, pitch, and roll), and detector acquisition information.  Ground control points consisting of small image chips with known geographic locations and elevations are matched to locations within the ASTER image.  Location differences between the satellite (systematic) information and the control point information are used to generate a mathematical model for adjusting the image to more closely agree with the control points.

What does the AST_L1T multi-file product consist of? What products are associated with the AST_L1T products?

The AST_L1T multi-file product consists of science data, and up to 2 Full Resolution geolocated Images(e.g. _V.tif, _T.tif). The Visual Geotiff image combines bands from the VNIR and SWIR telescopes to create a false color visual image. The Thermal Geotiff image combines 3 bands from the TIR telescope which gives a rough indication of temperature. The AST_L1T has 2 associated products created with each granule, the QA product (optional) and the Browse product. The associated QA product contains a quality assurance text file. The associated Browse product contains the low-resolution versions of the full resolution images, and if the QA product is created then a QA browse is created which indicates pictorially the quality of the visual bands. Each product has an associated metadata file.

Will ASTER Level 2 products be built from the AST_L1T input? Does the AST_L1T product replace any of the existing ASTER products?

Generation of ASTER Level 2 products from AST_L1T input is likely to be offered in the future, but a scheduled date has not been established.  At this time, the AST_L1T product does not “replace” existing ASTER products although there are similarities between the AST_L1B and AST_L1T scenes, specifically those with a ‘systematic’ level of correction. (Cross-reference Q7 of the Science Data Section listed above.)

Why isn't there an associated QA product for all AST_L1T products?

The QA product is only produced if precision correction is attempted and there is a GLS 2000 Standard Scene for the area the granule covers. The scene must be a day scene, not too cloudy, have a VNIR or SWIR telescope ON, and not be totally over water or too close to the poles.

Why does my tool show the scene in the Northern Hemisphere when it should be in the Southern Hemisphere?

Testing indicates that the tools can be expected to place the scene in the correct hemisphere on reference maps. The approach taken for AST_L1T mimics the approach used by Landsat; for all tools tested to date, the results provide consistent behavior for both products. In the AST_L1T HDF embedded metadata, Southern Hemisphere latitudes are stored as negative values for both degrees and UTM meters. In AST_L1T GeoTIF files, false northing is set to zero for both hemispheres. Consistent with this convention, the Southern Hemispheres will have negative values for the northing (y-axis). Certain tools (GDAL Info and GeoTIFF GUI specifically) will react to the Landsat convention by constructing projection coordinate system (projcs) character strings for reports that always use the “N” designator regardless of which hemisphere the scene resides. Thus the tools may indicate the scene in the Northern Hemisphere when it should be in the Southern Hemisphere. Testing indicates that the corner coordinates are always reported correctly and therefore place the scene in the correct hemisphere.

Why are there different Browse files associated with each AST_L1T scene?

The VNIR, TIR, and QA Browse files may not be present for all AST_L1T scenes since they are dependent on the VNIR and TIR telescopes being turned on and whether a geolocation verification is attempted. The individual browse files end in _BR_#.<browse type>.jpg where the # may be  2, 3, or 4, and the browse type may be VNIR, TIR or QA. The numbers represent the order the browse files were put into an HDF file for Ingest. The Browse files are put into the HDF file with VNIR first, TIR second and QA third. For granules that have all 3 Browse files the file names would end: _BR_2.VNIR.jpg _BR_3.TIR.jpg and _BR_4.QA.jpg. If the VNIR telescope is not on, the VNIR browse file is not created and the TIR file will end in _BR_2.TIR.jpg. If the VNIR telescope is on and the TIR telescope is off and the QA browse is created, then the QA browse file name will end with _BR_3.QA.jpg. 

The LP DAAC ASTER Level 1T (AST_L1T) terrain-corrected image product is available for immediate download to all users at no charge.
You can use any of the following methods to download the ASTER Level 1T (AST_L1T) data:

  • NASA Earthdata Search – enter ASTER in the Earthdata search and select ASTER Level 1 precision terrain corrected registered at-sensor radiance V003 data set to begin your search.
  • LP DAAC Data Pool – direct access to data directories for immediate HTTP retrieval is available via the ASTER collection.
  • USGS GloVis – under Collections, select ASTER, choose L1T Day (VNIR/SWIR/TIR) or L1T Night (SWIR/TIR) collection to find AST_L1T data at no cost.
  • USGS Earth Explorer – under Data Sets, select NASA LP DAAC Collections, ASTER Collections, ASTER Level 1 precision terrain corrected registered at-sensor radiance V003 data set to begin your search.