Processing MODIS Data

Creating Uncalibrated MODIS Datasets:

Datasets created by this method provide a quicklook for troubleshooting. They can also be used to create true-color images suitable for web pages.

modisin creates TeraScan MODIS datasets from CADU files on the pass disk or from PDS files. Output is a TeraScan data format (TDF) file that contains uncalibrated, earth-located MODIS channels. The data can be corrected for the bowtie effect caused by the MODIS sensor. The letter 'b' is appended to corrected output variables (for example, modis_ch03b). The whole pass can be processed or the data can be cut out to conform to a master file. Subsampling and other methods of limiting the size of the file are available to the user. The function requires an input pass number or PDS filename and an output TDF filename.

Suggested Processing Sequence:

Note: The functions in the steps below are links to the processing example.

Step [1]:	a.	Run modisin to process a pass on the pass disk or a PDS file. To process the whole pass and correct for the bowtie effect of the MODIS sensor, accept all the default responses to the prompts for input.
		Data in output TDF dataset is uncalibrated but earth-located.
		To limit the size of the output dataset, use master to create a file that defines an area of interest and map projection and then run modisin with use_master=yes.
		[Alternative methods for creating a master: TeraMaster (tmaster), master2, master4.]
	b.	Then view the contents of the TDF dataset with the contents function.
	b.	Note how the data is scaled to allow for dimension differences in the 250m, 500m, and 1000m channels. This allows all the variables to be stored in one dataset with the same earth location. The data is still in sensor scan projection.
	c.	To view the data in true color, run tvis -true & and load channels 1, 3, and 4. Suggested methods of combining and enhancing the channels for best RGB image (choose the one that works best for you):
		(1) In the Image Combine panel, select RGB (Red: modis_ch01, Green: modis_ch04, Blue: modis_ch03) and press Render Imagery. Then, enhance the combined image. (Suggestion: use a logarithmic enhancement.)
		(2) Enhance the channels individually, then use Image Combine to create an RGB composite (Red: modis_ch01, Green: modis_ch04, Blue: modis_ch03). Then do a box histogram enhancement of the RGB image by selecting a representative area that includes land, ocean, and clouds.
Step [2]:	a.	Run fastreg to register the data to the master.

Creating Calibrated MODIS Datasets:

Datasets created by this method are used for ocean color processing.

	pgs_modis_calib.csh uses IMAPP to process PDS files to calibrated radiances and convert them to level 1b earth-located TDF datasets. Output is a TeraScan data format (TDF) file that contains 36 calibrated and earth-located MODIS channels. The script requires an input pass number or PDS file. The format of the output TDF filename is t1.yyddd.hhmm.RADIANCE_tdf.
	modis_bt corrects for the bowtie effect caused by the MODIS sensor in datasets created by modis_calib.csh. The input TDF contains all 36 channels. The output TDF contains only the corrected channels, with a 'b' appended to the channel variable name. The function requires both input and output TDF filenames.
	get_modis_color.csh extracts the ocean-color channels (8, 9, 10, 11, 12, 13H, 15, and 16) from the data and renames them (ch1 through ch8) for input to modis_color for ocean color processing. The script requires an input TDF filename. The format of the output TDF filename is t1.yyddd.hhmm.clr_input.
	modis_color derives water-leaving radiances at 412, 443, 490, 510, and 555 nm from the data and writes them to the output dateset. (Other variables are written to the output dataset but are not currently being used: aerosols at 670 and 865 nm; CZCS pigments; chlorophyll a; K490; epsilon; aerosol optical thickness at 865 nm; NDVI; and processing flags.) The function requires both input and output TDF filenames.
	mod_chlor derives chlorophyll from the water-leaving radiances output by modis_color. The function requires both input and output TDF filenames.

Notes:

Before running pgs_modis_calib.csh:

Make sure there is enough available space for the output files in the directories defined in the script. Allow at least 7GB of space for the output HDF files, the final TDF, and intermediate output files.
Remove all pre-existing log files.
Make sure orbital elements are up to date. If you are processing an older pass, make sure the orbital elements for that date are available (view the file $SATDATA/terra-1/orbinput).

Suggested Processing Sequence:

Note: The functions in the steps below are links to the processing example.

Step [1]:	a.	Run pgs_modis_calib to use IMAPP to process a PDS file.
	a.	Scales and offsets from the calibrated HDF files are used to convert the data from scaled integer to radiance units in the output TDF dataset.
	b.	Then view the contents of the TDF dataset with the contents function.
	b.	Note how the data is scaled to allow for dimension differences in the 250m, 500m, and 1000m channels. This allows all the variables to be stored in one dataset with the same earth location.
Step [2]:	a.	Run modis_bt to correct the data for the bowtie effect caused by the MODIS sensor.
	a.	Correct channels 1-16 (1^16).
	b.	Then view the contents of the bowtie-corrected dataset. Channel variables will have a 'b' appended to their names.
	c.	Run navbox2 and nav2 to autonavigate the data in the output TDF dataset. Use the optional parameters use_pitch=yes and use_yaw=yes on the nav2 command line. Note: If your data cannot be navigated automatically, click here for details of manual navigation.
Step [3]:	[OPTIONAL] If disk space and processing time are important considerations for processing:
	a.	Use satmaster to create a file that defines an area of interest. (Use ephem to determine the new start time of the pass.)
	b.	Run mastersub to cut out the area of interest from the data. The data remains in sensor scan projection.
Step [4]:	a.	Use get_modis_color.csh to extract the ocean color channels (8 through 12, 13H, 15, and 16) from the calibrated dataset. The script renames the variables (ch1 through ch8) to make them ready for ocean-color processing.
	b.	Then, run modis_color on the output dataset to create an ocean color dataset with water-leaving radiance variables.
	c.	Finally, run mod_chlor on the ocean color dataset to derive a chlorophyll product from the water-leaving radiance.
Step [5]:	a.	Use master to create a file that defines an area of interest and map projection.
		[Alternative methods for creating a master: TeraMaster (tmaster), master2, master4.]
	b.	Run fastreg on the output from mod_chlor to register the data to the master.
	c.	Finally, run lsmask to create a landmask based on the master and then use emath2 to mask out areas of land in the data.

Processing MODIS Data

Creating Uncalibrated MODIS Datasets:

Suggested Processing Sequence:

Creating Calibrated MODIS Datasets:

Notes:

Suggested Processing Sequence:

SEE ALSO: