Earthdata

Matching images of Typhoon Rusa from the summer of 2002, showing the bow-tie effect (left) and geometric correction (right).

Bow-tie effect
Satellite sensors collect data by pixels. A pixel of data represents the electromagnetic energy reflected from or radiated by a given area of ground. All these pixels are like individual boxes in a grid, and they are put together to make the complete image. In a single rotation of its scan mirror, MODIS captures an area on Earth about 2300 kilometers wide by 10 kilometers tall (imagine a long, thin rectangle). An image is put together by stitching one scan–one strip–on to the next.

Stretch
Because the area is so large, MODIS is looking straight down at the Earth in the center, but is looking off to the side at an angle. Just like the spot illuminated by a flashlight becomes wider the farther away the beam is pointed, the pixels that make up the MODIS image get wider the farther they are from the center of the image. Each MODIS scan gets stretched out at the edges, making the shape of each scan–each strip of the image–look more like a bow-tie than a true rectangle.

Overlap
Because the strips are wider at the edge than the center, the scans don’t overlap in the center of the image, but they do overlap near the edges of the swath. This overlap creates a “double-vision” effect on the edges of the swath, as each point of the Earth’s surface appears in two adjacent scans. This geometric distortion also makes the line where scans come together more pronounced, essentially magnifying tiny differences between one side of the scan mirror and the other, especially at the edge of the swath.

Distortion
The real-time images are displayed as they are scanned by MODIS, without correction for geometric distortion. Since the observed pixels at the edge of the swath are bigger than they are in the center, they appear to be scrunched into a too-small grid box when displayed “as-is” in the real-time imagery. We select certain images for geometric correction, and hand tailor them for our image gallery.

How do you correct for the bow-tie effect and distortion?
The code we use to correct for the bow-tie effect and distortion is not documented for distribution, but there are several free tools that will give similar results. We recommend the MODIS Swath to Grid Toolkit (MS2GT), HDFLook, or HDF-EOS To GeoTIFF Conversion Tool (HEG).

How do I make true-color MODIS images?

To help you create your own true color MODIS images, we provide a tutorial (PDF) that tells you where to get the software you need, including a version of our corrected reflectance, fire detection, and vegetation index algorithms. Because of regulations on code distribution, code distribution is handled by the Direct Readout Laboratory.

Other alternatives for geometrically correcting MODIS data include MODIS Swath to Grid Toolkit (MS2GT), HDFLook, and HDF-EOS To GeoTIFF Conversion Tool (HEG).

I’ve tried making my own MODIS images from MOD09 data ordered from LAADS Web, but they don’t look like yours. Why not?

We make our imagery directly from the MODIS L1B data (the calibrated, geolocated radiances), not the standard MODIS surface reflectance product (MOD09). Our images are based on custom corrected reflectance code that has been developed with the specific aim of maintaining image integrity, as opposed to the standard MODIS surface reflectance product, which was developed with scientific integrity as the primary objective. For example, we do not use an atmospheric correction. In some cases, our images undergo additional hand-tweaking to enhance their appearance; for example, we may apply different color stretchs depending on the type of phenomenon (e.g. dust, phytoplankton bloom, clouds) we are highlighting.

Can you provide the full data file for an image I like in your gallery?

For data appearing on the Near Real-Time (Orbit Swath) Images page, links are provided to the near real-time full spectral resolution files HDF files on the LANCE MODIS FTP site. Links are also provided to the science quality full spectral resolution files HDF files on the LAADS Web site.

Can you send me geometrically corrected images and fire detections for my area of interest each day?

A substantial investment of time and resources is required to provide hand-tailored image processing for specific regions on a daily basis. The Rapid Response System provides these products daily to a small number of partner organizations, including the USDA Forest Service and Global Observation of Forest Cover and Land Dynamics (GOFC-GOLD) partners. These images are available on the MODIS Subsets page. As time and resources permit, we can occasionally respond to requests for coverage of significant events, and we are always interested in discussing the potential for meaningful scientific collaborations. If you are interested in proposing such a collaboration, please contact Earthdata Support.

Do you produce fire detections at night?

Nighttime fire detections are currently being collected for land regions across the world and are available from the Fire Information for Resource Management System (FIRMS).

What is the Julian day?

The Julian day, as we mean it, is the day of year (from 1 to 365 or 366). Most remote sensing data sets are identified using this Julian date rather than the calendar date, because it is a more straightforward time scale for automatic processes. Note this is different from the common meaning of Julian Day as in Julian calendar.

Where do you get the data used to draw political boundaries and coastlines on your gallery images?

This vector information comes from the National Geospatial-Intelligence Agency’s (NGA) Vector Map (VMap) Level 0. It is available from the NGA web site.

Where can I find the georeferencing information for your subset and gallery images?

The georeferencing information is contained in two places, the More information page and the JPEG worldfile. They are accessible from links on each image page.

More information
This page describes the projection used to georeference the image, the latitude and longitude of the image center and corners, and some additional information about the image. If the image is rotated, the “rotation angle” field will be given. Counterclockwise rotation is positive; the image is first projected then rotated. If the pixel sizes are not equal in the X and Y direction. the “x scale factor” will be given; it can also be calculated: x_scale_factor = y_pixel_size / x_pixel_size.

Note that the coordinates given for the image corners are for the outside corner of the pixel, and thus will differ by one half of the pixel size from the numbers given in the JPEG World File.

JPEG worldfile
This file has the file extension .jgw and is an ASCII text file consisting of 6 lines, each containing a single number:

Value Parameter Description 250.000000000000 A Pixel size in X direction 0.000000000000 D Rotation term for Y 0.000000000000 B Rotation term for X -250.000000000000 E Negative of pixel size in Y direction -692125.000000000000 C X location of center of upper left pixel 544375.000000000000 F Y location of center of upper left pixel

The following equations will calculate the map coordinates of the center of pixel (x,y) when the coordinates of the upper left pixel are (0,0):

Map_X = Ax + By + C

Map_Y = Dx + Ey + F

Note that the pixel size in the Y direction is negative because the map origin is at the lower left and the image origin is at the upper left. The pixel size is in degrees for the Plate Carree projection and in meters for all other projections.

Some GIS and Image Processing software packages will not correctly handle images whose rotation terms (the second and third lines) are not 0.0.

How do I use your subset and gallery images in my GIS or Image Processing software package?

We are only able to do limited testing with subset and gallery images using GIS and Image Processing software packages and would greatly appreciate any questions, comments, corrections, clarifications, etc. on these instructions from our users. Please send them to Earthdata Support.

The “Download JPG image with ancillary files (.zip)” link on the individual subset image pages produces a zip file which includes 4 files – the subset image (.jpg), a worldfile with coordinate information (.jgw), a file with projection information (.prj), and a file with the projection information wrapped in XML tags for ArcGIS (.jpg.aux.xml). It is hoped that these ancillary files will allow the subset images to be loaded into GIS and image processing software packages without the user having to manually set the projection information. Again, feedback and corrections would be greatly appreciated. (This zip file is not yet available for Gallery images.)

NOTE: The global fire maps do not have georeferencing information.

NOTE: When downloading the worldfile (it is a .jgw file) on Windows computers, some versions of Windows append a “.txt” to the “.jgw” when you download it with “Save as type:” set to “Text Document.” If the worldfile is saved with a “.txt” extension, the GIS program will not be able to find the worldfile. The name of the worldfile must match the name of the image file except with “.jgw” in place of “.jpg.” To prevent this problem when downloading the worldfile in a Windows environment, change “Save as type”: to “All Files”. Also be aware that, when checking the worldfile name using Windows Explorer, in some versions you can’t see the “.txt” if the “Hide extensions for known file types” box is checked in Folder Options. In this case, the file name looks correct when it is not.

ArcMap

To use subset and gallery images in ArcMap:

• First download the image (it is a .jpg file) and the worldfile (it is a .jgw file). You will also need to have some of the information from the “Display metadata” link handy. The “Display metadata” link is above the image on the left.
• Either before or after opening the image in ArcMap, you must set the Layer Coordinate System. The projection is specified by the “projection” in the “Display metadata” page.
  • Plate Carree
    • NOTE: Images in mid to high latitudes will appear wider than on our web site because our pixels are narrowed to approximate equal area while ArcMap forces the pixels to be square. (The “x scale factor” can be found in the “Display metadata” page, the x_pixel_size is the first line of the .jgw file, the y_pixel_size is the fourth line of the .jgw file (with a “-” prepended) and x_scale_factor = y_pixel_size / x_pixel_size)
    • Right-click on “Layers”, select “Properties…” then the “Coordinate System” tab
    • Select “Predefined” in the “Select a Coordinate System” box, then “Geographic Coordinate Systems” -> “World” -> WGS_1984
    • Click “OK” for “Data Frame Properties”
  • Lambert Azimuthal – Sinusoidal – Stereographic
    • Right-click on “Layers”, select “Properties…” then the “Coordinate System” tab
    • Select “<custom>” in the “Select a Coordinate System” box
    • Click the “New” button then select “Projected Coordinate System…”
    • Enter a name in the “Name” box
    • Projection-specific instructions:
      • Lambert Azimuthal
        • In the “Projection” panel, under “Name” select “Lambert_Azimutal_Equal_Area”
        • Set “False_Easting” and “False_Northing” to 0
        • Set “Central_Meridian” to the value from the “projection center lon” in the “Display metadata” page
        • Set “Latitude_Of_Origin” to the value from the “projection center lat” in the “Display metadata” page
      • Sinusoidal
        • In the “Projection” panel, under “Name” select “Sinusoidal”
        • Set “False_Easting” and “False_Northing” to 0
        • Set “Central_Meridian” to the value from the “projection center lon” in the “Display metadata” page
      • Stereographic using Stereographic_South(North)_Pole
        • In the “Projection” panel, under “Name” select “Stereographic_South(North)_Pole”
        • Set “False_Easting” and “False_Northing” to 0
        • Set “Central_Meridian” to the value from the “projection center lon” in the “Display metadata” page
        • Set “Standard_Parallel_1″ to the value from the “standard parallel” in the “Display metadata” page
      • Stereographic using Stereographic
        • In the “Projection” panel, under “Name” select “Stereographic”
        • Set “False_Easting” and “False_Northing” to 0
        • Set “Central_Meridian” to the value from the “projection center lon” in the “Display metadata” page
        • Set “Scale_Factor” to the value from the “scale at natural origin” in the “Display metadata” page
        • Set “Latitude_of_Origin” to the value from the “projection center lat” in the “Display metadata” page
    • In the “Linear Unit” panel, under “Name” select “Meter”
    • If if the value of “ellipsoid” is “WGS84″ in the “Display metadata” page, then in the “Geographic Coordinate System” panel, click the “Select” button, then “World” then “WGS84.prj” (Now skip to the “Click OK for New Projected Coordinate System” step)
    • If if the value of “ellipsoid” is “Sphere” in the “Display metadata” page, then in the “Geographic Coordinate System” panel, click the “New…” button
    • Enter a name in the “Name” box
    • In the “Datum” panel, under “Name” select “<custom>”
    • In the “Spheroid” panel, under “Name” select “<custom>”
    • Set both “Semimajor Axis” and “Semiminor Axis” to the value from the “Earth radius (km)” in the “Display metadata” page. You will need to move the decimal point 3 digits to the right to convert from kilometers to meters.
    • In the “Angular Units” panel, under “Name” select “Degree”
    • In the “Prime Meridian” panel, under “Name” select “Greenwich”
    • Click “Finish” for “New Geographic Coordinate System”
    • Click “Finish” for “New Projected Coordinate System”
    • Click “OK” for “Data Frame Properties”

ArcView

NOTE: Some gallery images cannot be used in ArcView because they are rotated in order to capture more of the scene in a rectangular image. You can determine if an image is rotated or not by checking for the existance of the “rotation angle” keyword under the “Display metadata” link for that image.

To use subset and unrotated gallery images in ArcView:

• First download the image (it is a .jpg file) and the worldfile (it is a .jgw file). You will have to replace all the periods in the file names (“.”), except the final one before the extension, with underscores (“_”). You can do this during the download process or afterwards. (Or you can completely rename the files, making sure that the portion before the .jpg and .jgw are identical.) You will also need to have some of the information from the “Display metadata” link handy. The “Display metadata” link is above the image on the left.
• Make sure the ArcView JPEG Extension is enabled by going to File->Extension and checking “JPEG (JFIF) Image Support”.
• Either before or after opening the image in ArcView, you must set the View Projection. The projection is specified by the “projection” in the “Display metadata” page.
  • Plate Carree
    • NOTE: (1) The settings below will most likely be the default settings. (2) Images in mid to high latitudes will appear wider than on our web site because our pixels are narrowed to approximate equal area while ArcView forces the pixels to be square. (The “x scale factor” can be found in the “Display metadata” page, the x_pixel_size is the first line of the .jgw file, the y_pixel_size is the fourth line of the .jgw file (with a “-” prepended) and x_scale_factor = y_pixel_size / x_pixel_size)
    • Go to View->Properties
    • Set “Map Units” to “decimal degrees”
    • Click the “Projection…” button
    • Click the “Standard” button
    • Set “Category” to “Projections of the World”"
    • Set “Type” to “Geographic”
  • Lambert Azimuthal
    • Go to View->Properties
    • Set “Map Units” to “meters”
    • Click the “Projection…” button
    • Click the “Custom” button
    • Set “Projection” to “Lambert Equal-Area Azimuthal”
    • “Spheroid” will be forced to a value of “Sphere”
    • Set “Central Meridian” to the value from the “projection center lon” in the “Display metadata” page
    • Set “Reference Latitude” to the value from the “projection center lat” in the “Display metadata” page
  • Sinusoidal
    • Go to View->Properties
    • Set “Map Units” to “meters”
    • Click the “Projection…” button
    • Click the “Custom” button
    • Set “Projection” to “Sinusoidal”
    • “Spheroid” will be forced to a value of “Sphere”
    • Set “Central Meridian” to the value from the “projection center lon” in the “Display metadata” page

ENVI

NOTE: Some gallery images cannot be used in ENVI because they are rotated in order to capture more of the scene in a rectangular image. You can determine if an image is rotated or not by checking for the existance of the “rotation angle” keyword under the “Display metadata” link for that image.

To use subset and unrotated gallery images in ENVI:

• First download the image (it is a .jpg file) and the worldfile (it is a .jgw file). You will also need to have some of the information from the “Display metadata” link handy. The “Display metadata” link is above the image on the left.
• In order to save the new projections you will create, for projections other than Plate Carree, you will need to have a writeable copy of the file “map_proj.txt”. First make a copy of the original ENVI file in your own directory space. (In our SGI installation, the file was found in the directory /usr/local/lib/idl_5.5/products/envi_3.5/map_proj.) Second, you must make this copy of “map_proj.txt” your default copy by selecting File -> Preferences -> User Defined Files… and using “Choose” to select “Default Map Projection File”. You must restart ENVI for the change to take effect.
• Due to a bug (in some ENVI versions) in specifying some parameters for Lambert projections, you may need to create the projection in the “map_proj.txt” file. Instructions are provided below.
• The projection is specified by the “projection” in the “Display metadata” page.
  • Plate Carree
    • NOTE: Pixels in images from mid to high latitudes will not be square in degrees because they are narrowed to approximate equal area. (The “x scale factor” can be found in the “Display metadata” page, the x_pixel_size is the first line of the .jgw file, the y_pixel_size is the fourth line of the .jgw file (with a “-” prepended) and x_scale_factor = y_pixel_size / x_pixel_size)
    • Open the JPEG file with Open External File -> Generic Formats -> JPEG. Select the filename and click “OK”
    • The “JPEG World File Input Projection” dialog will appear.
    • In the “Select Input projection” box, select “Geographic Lat/Lon”
    • In the “Datum…” box, select “WGS-84″
    • In the “Units…” box, select “Degrees”
    • Click “OK” for “JPEG World File Input Projection”
  • Lambert Azimuthal (using “map_proj.txt” file)
    • Edit your “map_proj.txt” file and add a new line containing the following:
    • A “36″ for Lambert Azimuthal Equal Area (sphere), followed by a comma
    • The value from the “Earth radius (km)” in the “Display metadata” page, followed by a comma. You will need to move the decimal point 3 digits to the right to convert from kilometers to meters.
    • The value from the “projection center lat” in the “Display metadata” page, followed by a comma
    • The value from the “projection center lon” in the “Display metadata” page, followed by a comma
    • Finally, “0.0,0.0,” (for false easting and northing) followed by a unique name for the projection
    • Save the file with the new projection definition and restart ENVI for the change to take effect.
    • Open the JPEG file with Open External File -> Generic Formats -> JPEG. Select the filename and click “OK”
    • The “JPEG World File Input Projection” dialog will appear.
    • In the “Select Input projection” box, select the name of your new projection
    • In the “Datum…” box, select “<none>”
    • In the “Units…” box, select “Meters”
    • Click “OK” for “JPEG World File Input Projection”
  • Lambert Azimuthal (using “Customized Map Projection Definition”)
    • Open the JPEG file with Open External File -> Generic Formats -> JPEG. Select the filename and click “OK”
    • The “JPEG World File Input Projection” dialog will appear.
    • In the “Select Input projection” box, select “New…”
    • Enter a unique name in the “Projection Name” box
    • In the “Projection Type” box, select “Azimuthal Equal Area (sphere)”
    • Set “Sphere Radius” to the value from the “Earth radius (km)” in the “Display metadata” page. You will need to move the decimal point 3 digits to the right to convert from kilometers to meters.
    • Set “False easting” and “False northing” to 0
    • Set “Latitude of projection origin” to the value from the “projection center lat” in the “Display metadata” page
    • Set “Longitude of central meridian” to the value from the “projection center lon” in the “Display metadata” page
    • Click “OK” for “Customized Map Projection Definition”
    • You will asked if you want to save the customized projection to the “map_proj.txt” and, if so, to select a filename and if you want to overwrite.
    • If you saved the projection, the next time you open the JPEG file, you can simply select the customized projection name from the “Select Input projection” box.
    • In the “Datum…” box, select “<none>”
    • In the “Units…” box, select “Meters”
    • Click “OK” for “JPEG World File Input Projection”
  • Sinusoidal
    • Open the JPEG file with Open External File -> Generic Formats -> JPEG. Select the filename and click “OK”
    • The “JPEG World File Input Projection” dialog will appear.
    • In the “Select Input projection” box, select “New…”
    • Enter a unique name in the “Projection Name” box
    • In the “Projection Type” box, select “Sinusoidal”
    • Set “Sphere Radius” to the value from the “Earth radius (km)” in the “Display metadata” page. You will need to move the decimal point 3 digits to the right to convert from kilometers to meters.
    • Set “False easting” and “False northing” to 0
    • Set “Longitude of central meridian” to the value from the “projection center lon” in the “Display metadata” page
    • Click “OK” for “Customized Map Projection Definition”
    • You will asked if you want to save the customized projection to the “map_proj.txt” and, if so, to select a filename and if you want to overwrite.
    • If you saved the projection, the next time you open the JPEG file, you can simply select the customized projection name from the “Select Input projection” box.
    • In the “Datum…” box, select “<none>”
    • In the “Units…” box, select “Meters”
    • Click “OK” for “JPEG World File Input Projection”

How do I use the “Search by keyword” feature in your gallery?

Each Rapid Response Gallery image is annotated with keywords to aid in finding images of interest. The keywords fall into two categories, geographic, such as country, US state, Canadian province, and water body names, and image subject, such as “fire”, “hurricane”, “phytoplankton bloom”, “volcano”, “dust”, “smoke”, “iceberg”, “snow”, etc.

This is a very simple search which searches on a literal string. The search is not case sensitive and the string can contain blanks (e.g. “new mexico”) or partial words. The search input can be multiple literal strings seperated by upper case “OR” (e.g., “hurricane OR typhoon OR tropical storm”). Similarly the search input can be a pair of strings separated by upper case “AND” (e.g., “canada AND fire”). The matching images are displayed with the most recent images at the top of the page. If no matches are found, only the search boxes will be displayed.

Some examples of keywords which will not work are city names and continent names. Continent names will return some images from that continent but it will not be a complete list.

How can I tell what time a subset image was acquired?

To find out when MODIS collected the data used in the subset image, click on “Display metadata (including time of input data)” above the image. Under “5 minute swath data used for this image”, there is a list of the MODIS granules that are included in the image. Click on the time of the overpass in UTC to see the real-time browse image for each granule.

The subset images may be composited from data collected in different MODIS overpasses at different times. Because the images cover set geographic regions, not a particular MODIS swath, the automated system that generates the subset images takes data from the overpasses that are closest to being directly overhead. Where overpasses overlap in the high latitudes of the north and the south, data from two or more overpasses (100 minutes apart) may be used to create the subset image. In these cases, a faint line may be visible where the two overpasses meet. White shows where data are not available either because MODIS has not imaged that region yet or because the area is out of the instrument’s range for that day. In equatorial regions, where overpasses do not overlap, MODIS data may not be available for the entire region every day. In these cases, a white wedge where no data were available may separate two MODIS overpasses.

How can I set up a direct broadcast system to receive data from the MODIS sensor?

The Direct Readout Laboratory can provide more information about setting up a direct broadcast system. You can also find additional information on the MODIS home page.

How do I convert to NDVI values from the colors in your NDVI images?

The colors used for the ndvi subset images (since June 2, 2004 (2004154), except NewMexico) can only be converted back to one-tenth NDVI ranges, as follows:

Color NDVI R G B 153 204 255 <=0.0 225 175 100 0.0-0.1 255 225 150 0.1-0.2 225 255 175 0.2-0.3 152 255 152 0.3-0.4 102 255 102 0.4-0.5 51 204 51 0.5-0.6 0 153 0 0.6-0.7 0 102 0 >0.7 255 255 255 no data

For realtime images, subset images before June 2, 2004 (2004154), and the New Mexico subset, use the following conversion table:

Color NDVI R G B 140 190 255 <=0.0 (includes missing data) 255 245 225 0.00-0.05 255 225 200 0.05-0.10 245 201 140 0.10-0.15 255 221 85 0.15-0.20 235 190 55 0.20-0.25 250 255 180 0.25-0.30 230 250 155 0.30-0.35 205 255 105 0.35-0.40 175 240 90 0.40-0.45 160 245 165 0.45-0.50 130 225 135 0.50-0.55 120 200 120 0.55-0.60 158 198 108 0.60-0.65 140 175 70 0.65-0.70 70 185 40 0.70-0.75 50 150 20 0.75-0.80 20 120 80 0.80-0.85 30 80 0 0.85-0.90 120 20 20 >0.90