Skip to content. | Skip to navigation

Personal tools
You are here: GES DISC Home Services Data Tools simap SIMAP Readme

SIMAP Readme


New, Version 4, SIMAP is posted.Major improvements impact the way simap handles the internal geolocations of all MODIS Level 1B products, including the institutional Direct Broadcast products.Another big upgrade is the inclusion of patches that allow certain AIRS Level 2 and 1B products to be mapped. In better details, the changes introduced with SIMAP.4 are:
1. Introduced special inter- and extrapolation of geolocations for MODIS L1B, including direct broadcast. Thus, the georeferencing (mapping) is substantially improved, especially in the case of 1-km L1B and internal geolocation arrays.
2. Introduced dynamic offsets,ia,ja, to better fill holes in the mapped images.
3.Made changes allowing certain AIRS data to be mapped as well, including the visible L1B, AIRVBRAD. SIMAP can now stitch that AIRS data.
4. The CONVERT_COORD is vectorized to speed up the process.If the geolocation arrays are too big, the code will switch to a FOR-cycle to avoid possible memory problems.


Added a function SCALEGEO to scale integer geolocations to real values, if local attributes like "Slope" and "Intercept" to the geolocation arrays are present. Such is the case with MODOCQC/MYDOCQC, and the joint atmosphere product, MODATML2/MYDATML2.

simap also has a slight change to the common block, that makes more variables be seen from the batch file.This was done to allow more sophisticated map generations where more than three layers can be superimposed, e.g. to combine true color visible image with some science parameter.This, however, requires advanced knowledge of MODIS quality flags, science parameters valid ranges, and logical operators in IDL. This job can be quite different from parameter to parameters. Thus, we can provide sample batch file on request which may not match you needs though, and regret to not being able to support it further.


More and more users have newer versions of IDL where the GIF license was not acquired by RSI.IDL5.3 had both, GIF and PNG, but needed the image flipped over before saving to the PNG file. This flipping is not neccesary in IDL5.5.All these annoying differences made us introduce a check of the IDL version, before writing a one-layered image to a GIF or PNG file.If the IDL version is 5.3 or older, then the program will use GIF interface. For newer versions (IDL5.5) the program will use the PNG instead.Image examples here, however, are gif files produced by IDL 5.3, and are intentionally left for users with older browsers.


    The tool described here is an IDL based script that is designed to read and map MODIS level 1B (L1B)  and level 2 (L2 ) Ocean and Atmosphere products.  The main intention is to have a non-interactive, command-line-executed tool, that can generate various maps of the above products.  The resulting maps are scaled to physical units (radiances, concentrations, brightness temperatures, etc.) and are saved in binary files.  At the same time image files are also saved, and their type depends on the task performed.  A main feature is composition of global maps, or stitching of multiple L1B or L2 granules on a single world map.  This can be useful in e.g. momentary display of Direct Broadcast data, or producing  high resolution continental maps of US.  Another useful feature is producing of large galleries of images from a large amount of granules.  This can be used to quickly and automatically produce browse images from arbitrary amount of granules and parameters, instead of doing images manually one by one.  The feature can be used to produce movies from time series of MODIS L1B and L2 granules, or for quality control of big volumes of data.  A feature is also a module in the main code that detects invalid geolocations, and patches them by means of  an interpolation that uses neighboring correct values.
    To repeat, the design of the tool is intended for automatic production systems (like S4P, for example), where mapped remote sensing products are desired (GIS for example).

    IDL 5.3 or higher version is required to be installed on the machine where the script will run.  When working on MODIS 250m resolution data (MOD02QKM) a good memory resource must be provided, in general more than 500 MB.  The script by its self does not need installation.  There are only two text files that experienced IDL users can even mold to their needs.  Thus, it is platform independent, as long as the platform has the IDL 5.3 or higher.  However, the batch mode design of the script is intended to take the full benefit ofa UNIX system (inclusion in shell or Perl scripts, running in cron jobs, etc.)

    The tool consists of just two files.   One is an IDL batch file, and the other is the main code,  The batch file contains configuration switches and all needed information to run the main code.  The batch file consists of simple commands that you would normally type in the IDL command line.  In the batch file, after a setup section, the main code is executed by a simple single line instruction.  In general, the main code does not need any intervention.  Yet, there are things that user may want to change, and these will be explained.  The batch file format,
   (execute code)
can be repeated arbitrary number of times in a single batch file thus giving the user a lot of choices of batch jobs.   E.g. once the main setup has been made, a gallery of [mapped] 40 Ocean L2 parameters can be generated, switching only  linear/logarithmic scales, color scales, or image intensities, for example.  These 40 images can be single granules, or 40 continental maps made from tens of granules.
     You may want to place the main code in one place, and then to keep different batch files for different batch jobs in separate directories.  All of them can use just that one main code,  However, it is your responsibility to ensure that IDL 'sees' needed directories.  Setting $PATH in UNIX, e.g., won't help!   Here is what you can do.

      First, regarding the main code, it is compiled from the batch file, and you have to ensure that the main code, , is visible to the IDL compiler.  There are several ways of doing that, choose one:
      i) The simplest way is to hard code the path to the main code in you batch files in the .compile line (see example batch files below):
          .compile /path/
      ii) Add a line before the .compile line, in the batch files (see example batch files below) that says and uses the right separator for your operating system (e.g. ':' for UNIX, ';' for Windows):
      iii) Add a line before the .compile line in the batch files (see example batch file below) that says (don't worry about the right separator)
      iv) Keep batch files and  the main code in one directory, make this directory current, and run IDL from the same directory.
      Second, regarding the batch file.IT IS THE BATCH FILE that executes the program.Thus, if running from an IDL command line, then type and execute (idl> is the IDL prompt):
where "path" is the name of directory where you keep your IDL batch file, batchfile.bat (see example batch files below).   If running from a UNIX box, type and execute (% is the UNIX shell prompt):
%  idl   /path/batchfile.bat

    You don't need to be a programmer to put together your own batch file, given the examples we provide here.  The switches and parameters of the batchfile will define in what mode and how the main code will execute.  Be very careful when choosing your setup, as it will determine how the program will run, if run at all.  Once you set all switches and parameters in the batch file for a given job, you don't need to reset them if you do same job over many granules.  The main code, as we pointed out, does not need intervention.
    4.1.  simapL1B_1km.bat
    The following is an explanation of all switches and parameters in the example simapL1B_1km.bat  we provide.  More batch examples for many particular cases are provided afterwards.  simapL1B_1km.bat is a batch file setup that makes a gallery of gray scale mapped images from MODIS land bands (1,4,3) at 1-km resolution.  The result is automatically saved in png and in binary "dat" files.  In the png files, all components of a scientific graphics are included: axis box, labels, gray scale legend, name of the parameter and units (taken from the HDF attributes).  In the binaries, only what you see inside the map is saved in physical units and is available for further processing.  So, the mapped data is available in image and binary files.  Let's go through this batch file line by line.  We will remind again, once you set all switches and parameters in the batch file, you don't need to reset them if you do same job over many granules.

.compile simap
The first line makes a full reset of the IDL state, and the second is compiling the main code.
These two lines do not need to be repeated any longer in the batch file. However, they need to be before any other commands and be in that order. Note, that IDL must see where the main code, simap, resides. We mentioned how to do it in the section "HOW TO RUN" above. The simplest way is just to hard code the path: .compile /path/

These two lines initialize a data structure that will keep the names of Scientific Data Sets (SDS) and channels, given SDS are three-dimesional, you want to retrieve.  In the rare case you want to make more than 20 selections, change this number.  Otherwise, no changes needed here.  Note, it is actually an array of structures.  The elements are sel[0], sel[1] and so forth.  Each element has (or points) three thingies: "name", "chans" and "numch".  You don't need to worry about "numch".  How to set "name" and "chans" will be described shortly.

inpath= '/ftp/data/modis/samples/ radiometric_geolocation/L1B/MOD021KM'
These are the path to your input HDF files, and the path to output the mapped data.  I.e. image and binary files will go to the directory stored in outpath.  These may be changed any time before the main code is executed.  Thus for every run of the main code you can take data from, and distribute to, different locations.

This is the name of your input HDF file.  It can be a single HDF file name, or a text file containing the list of granules (HDF files) you want to process.  In the latter case, the name must be literally "filelist", and so hdfname="filelist".  The file "filelist" must contain an HDF name per line.  It may contain just one name.  Paths can be included, but then you must set inpath to a null string:  inpath="".   This is NOT a good practice, though, and we discourage from that.  The reason is, if you ask for external geolocations, then the code will search in the "inpath" directory and will fail, because it expects external geolocations and the data in the same directory.


These are the SDS selections, and channels names for the corresponding SDS, you make for the following run of the main code.  All of these need to be quoted and comma separated within the brackets!  The reason is the MODIS L1B 1km data: it has band names '13lo', '13hi', '14lo', '14hi' that can only be entered as strings.  The SDS EV_1km_RefSB has band names starting from '8', going through '13lo', '13hi', '14lo', '14hi', and the last two are '19' and '26'.   This is why the code expects band names, and not consecutive indexes of the third dimension, for L1B data!  All other data when needed, e.g. Atmosphere L2 profiles, can be selected as indexes of the third dimension.  You will see detailed examples below.

;============FLAGS DEFINITIONS==================

If set to 1, this switch will activate an interpolation that will attempt to fill up artifacts of mapping procedure.  Such artifacts, seen as 'holes', will appear when the requested image size is comparable to or larger than the SDS array size. Almost certainly you will encounter this case when working with MODIS Atmosphere L2 products.  Set to 0 to turn off this interpolation.When stitching granules (global maps), it turns off automatically.

Set this switch to 1 to use EXTERNAL geolocations from MOD03 product.  Set to 0 to use INTERNAL geolocations.  When mapping MODIS Ocean products, you must supply corresponding MOD03 file and set extgeo=1.  You may choose to use external geolocation when mapping L1B and Atmosphere L2 as well, but for these cases this is not necessary.

Set to 1 to extrapolate geolocations to the size of the SDS.  It must be 1 when using internal geolocations and L1B data.  It is a good idea to keep it always 1.  You may set igeo=0 when using external geolocations with L1B 1km or Ocean L2.  When doing large jobs in these cases it will be beneficial to set igeo=0, because geolocation arrays size is the same SDS and there is no need to extrapolate.  However, setting igeo=1 will save you from thinking all the time of relative array sizes.

This is a switch to list SDS and print them on the current output device.  If you ever find use of it, here are the states:
0=no list
1=short list, SDS names only, no attributes
2=long list, SDS names, attributes and their values

This is a flag for the type of image to be generated after all procedures
are completed.
Set to 3 to show image composite from the first 3 channels selected.
Set to 1 to show b/w first layer
Set to 0 not to show image on the monitor
Note: when doing global maps (see gallery=0), the resulting image will be displayed only if you run in debug mode (see debug=-1) and after all granules are placed on the global map.  When doing galleries of images this flag will automatically turn to 0 to save time.  You may run a job for a large gallery unattended, and then review it with some png- or jpg- viewer.

This is a switch to fill continents with a uniform gray color.  It is useful for Ocean L2 products.
Set to 1 to fill continents
 Set to 0 not to fill continents

If 1, this flag will cause map limits (boundaries) to be automatically set.
Set to 0 to use the default map limits that you will set.  See map_limit below.

Set to 1 to automatically set image brightness.
Set to 0 to use default image brightness ranges that you will set. See "intens" below.

This flag causes the image to be presented in log scale.
Set to 1 to show image in log scale.  This is useful for certain Ocean L2 parameters.   This will not affect the data stored in the binary files, only the images.

This flag regulates whether a global map will be produced from a list of granules, (stitching mode), or a gallery of  image files. When this flag is set to 1, the code will generate galleries of image files.
0=no gallery, will do global map (stitching mode)
1=png or jpeg gallery of gray/color scale, of true color, RGB, image files. If  True_col=1 then grey/color scale.  If True_col=3 then the gallery will consist of RGB images composed from the first three selections.  See below for the ctab parameter to understand how to set color table.

This flag regulates the debug mode.  When debug=-1 all processing messages, including errors, will print to the monitor.  Set to 4 to print to a log file.  The name of the log file will be  log.txt and currently is set to be in the directory set by 'outpath'

;==========PARAMETERS DEFINITIONS=============

 xstart="";leave as NULL string for full size
 xend="";or give numerical value
This section allows for a spatial subset by taking a row/column subset.The quotes are an empty string that signals that no subset is needed.To initiate a subset, replace any one (or all) of the empty strings with a numerical value. xstart/xend correspond to the along scan dimension ("frame numbers"), while the ystart/yend correspond to the along track dimension ("line numbers").

 ;---------Default map settings------------  
These are default, hard coded, map boundaries.  They will be active if  automap=0.  If automap=1, they are ignored and the map boundaries are automatically determined from geolocation arrays. In case automap=1 and polar regions (DAYNIGHTFLAG is "Both"), a special procedure is used to form an 8-element vector instead. The valid range for latitudes is +-90 deg, and for longitudes is +-180 deg.

The latitude of the point on the Earth's surface to be mapped to the center of the projection plane.  Setting it to 0 will give an easy to handle rectangular grid.  However, it results in large distortions at higher latitudes.  If such distortions are an issue, try something like center_lat=latmin+(latmax-latmin)/2

 center_lon=lonmin+ (lonmax-lonmin)/2.
The longitude of the point on the Earth's surface to be mapped to the center of the projection plane.

This defines map position and size relative to the graphic window.

This will define the type of projection.  Two are available so far, but more can be added easily.
1 = cylindrical
2 = stereo

 intens[*,0]=[5.,80] ;red
 intens[*,1]=[15.,80] ;green
 intens[*,2]=[40.,100] ;blue
These are default image birightness ranges.  They will be active if autointen=0.

For 3D arrays, this is the dimension to fix.  In IDL, arrays dimensions are e.g. array[x,y,z]. Set slice=1 or 2 or 3 to fix
 x or y or z dimension, correspondingly. For most SDS's slice=3, i.e. they are stacked over the z-dimension. However, for example MODIS Atmosphere L2 Quality_Assurance needs slice=1 !!!  I.e. it is stacked over the x-dimension.

This is the size in pixels of the map to be produced.  NXM and NYM are the horizontal and vertical sizes, respectively.  Note: the size of the graphic window, that accommodates the map, is calculated from this input and is larger.  Thus, the array sizes stored in the binary files are NXM x NYM.  The binary files are always 4B (32 bit) FLOATS, because the data there are scaled to physical units.    In the case you make a multichannel selection and a global map, all channels are stacked in one binary file.   In the case you make a gallery, all channels are saved in a separate binary files.  Each file has the channel/parameter name embedded in the file name.  The file names are formed automatically, you don't need to worry about that.  Now, the size of the image file is the size of the graphic window.  The size of the graphic window is NX x NY and is calculated below, you don't need to change or set it.  Do not worry if the map is larger than your monitor.  E.g. if you monitor is 640x480 and the map is 2000x2000, you still will be able to see the map (pending on the mode) in the IDL environment because we use a sliding window.  Recall that in "gallery" mode graphics are not send to monitor but are saved in image files to speed up the operation.

This parameter holds the IDL color table number.  0 is linear black and white and is the default.  ctab=33 will call for a rainbow.  This parameter does not influence the binary files, only the image files that a single layer.  That is, ctab does not influence the composite RGB jpg files either.  See for tricks and tips below how to use it when choosing the right color table.

This is a background inside the map and outside the granule(s) in terms of color index.  In general, IDL has 256 color indexes in a given color table (this, however, depending on your monitor and terminal is not granted).  Thus, when the color table is ctab=33 (rainbow), then backgnd=128 will give the green background.

 ;this will be the size of the graphic window to
 ;hold the map:


Finally, this will execute the code with the settings from above.

Document Actions
NASA Logo -
NASA Privacy Policy and Important Notices
Last updated: Sep 10, 2010 02:20 PM ET