New Mexico Tech
Earth and Environmental Science

ERTH 455 / GEOP 555 - Geodetic Methods

Lab 8: Processing w/ gmtSAR: Timeseries; Getting SAR data

"I like my crust deformed."
UNAVCO bumper sticker

Note that you DO have to work on redoubt for some things today!

Introduction

This week you will process some Sentinel 1 data as a time series and learn something about where and how to get data you might be interested in. Ideally, we'd start you out on getting some data and then you'll process your own time series, you'll do some of that for the homework!

A note of warning: there's a lot of vocabulary flying (ha!) around in InSAR processing. Most of these are standard satellite related terms. If there's something you don't understand, it's usually a good idea to search for the term, the satellite and read up on it.

1. gmtSAR - time series example - Preparation

If things are unclear, refer to the gmtSAR documentation

Change to your working directory and link to the test data (no copying required):

    $> cd $GEOP555
    $> mkdir lab08
    $> cd lab08
    $> cp /data/GEOP555/lab08_make_links.sh ./
    $> ./lab08_make_links.sh
    $> ls
	

A directory listing should look similar to this:

	redoubt:/data/GEOP555/gps/lab08> ls
	lab08_make_links.sh  raw  topo
	

Preprocessing I: Check on Baselines

First we'll need some preprocessing and alignment of the SAR images. For this we'll use preproc_batch_tops.csh as we're dealing with Sentinel 1 data in TOPS mode. Running this command gives the following help:

	redoubt:/data/GEOP555/gps/lab08> preproc_batch_tops.csh 

	Usage: preproc_batch_tops.csh data.in dem.grd mode
	  preprocess and align a set of tops images in data.in, precise orbits required

	  format of data.in:
	    image_name:orbit_name

	  example of data.in
	    s1a-iw1-slc-vv-20150626...001:s1a-iw1-slc-vv-20150626...001:s1a-iw1-slc-vv-20150626...001:S1A_OPER_AUX_POEORB_V20150601_20150603.EOF
	    s1a-iw1-slc-vv-20150715...001:s1a-iw1-slc-vv-20150715...001:s1a-iw1-slc-vv-20150715...001:S1A_OPER_AUX_POEORB_V20150625_20150627.EOF

	  outputs:
	    baseline.ps align_table.ra (contains info for precise geomatric alignment)
	    *.PRM *.LED *.SLC(mode 2)

	  Note:
	    The names must be in time order in each line to be stitched together		
	

This means you'll need to create a mapping between measurements and orbits. Go into the raw directory and list its contents:

    $> cd raw
    $> ls
	

You'll get something like this:

	s1b-iw3-slc-vv-20170515t005941-20170515t010006-005602-009d02-006.tiff  s1b-iw3-slc-vv-20170831t005947-20170831t010012-007177-00ca6a-006.xml
	s1b-iw3-slc-vv-20170515t005941-20170515t010006-005602-009d02-006.xml   s1b-iw3-slc-vv-20170912t005948-20170912t010013-007352-00cf90-006.tiff
	s1b-iw3-slc-vv-20170527t005942-20170527t010007-005777-00a205-006.tiff  s1b-iw3-slc-vv-20170912t005948-20170912t010013-007352-00cf90-006.xml
	s1b-iw3-slc-vv-20170527t005942-20170527t010007-005777-00a205-006.xml   S1B_OPER_AUX_POEORB_OPOD_20170604T111305_V20170514T225942_20170516T005942.EOF
	s1b-iw3-slc-vv-20170608t005943-20170608t010008-005952-00a717-006.tiff  S1B_OPER_AUX_POEORB_OPOD_20170616T111329_V20170526T225942_20170528T005942.EOF
	s1b-iw3-slc-vv-20170608t005943-20170608t010008-005952-00a717-006.xml   S1B_OPER_AUX_POEORB_OPOD_20170622T111334_V20170601T225942_20170603T005942.EOF
	s1b-iw3-slc-vv-20170702t005944-20170702t010009-006302-00b14b-006.tiff  S1B_OPER_AUX_POEORB_OPOD_20170626T111333_V20170605T225942_20170607T005942.EOF
	s1b-iw3-slc-vv-20170702t005944-20170702t010009-006302-00b14b-006.xml   S1B_OPER_AUX_POEORB_OPOD_20170628T111321_V20170607T225942_20170609T005942.EOF
	s1b-iw3-slc-vv-20170714t005945-20170714t010010-006477-00b639-006.tiff  S1B_OPER_AUX_POEORB_OPOD_20170704T111629_V20170613T225942_20170615T005942.EOF
	s1b-iw3-slc-vv-20170714t005945-20170714t010010-006477-00b639-006.xml   S1B_OPER_AUX_POEORB_OPOD_20170716T111254_V20170625T225942_20170627T005942.EOF
	s1b-iw3-slc-vv-20170726t005945-20170726t010011-006652-00bb37-006.tiff  S1B_OPER_AUX_POEORB_OPOD_20170722T111441_V20170701T225942_20170703T005942.EOF
	s1b-iw3-slc-vv-20170726t005945-20170726t010011-006652-00bb37-006.xml   S1B_OPER_AUX_POEORB_OPOD_20170803T111617_V20170713T225942_20170715T005942.EOF
	s1b-iw3-slc-vv-20170807t005946-20170807t010011-006827-00c043-006.tiff  S1B_OPER_AUX_POEORB_OPOD_20170815T111344_V20170725T225942_20170727T005942.EOF
	s1b-iw3-slc-vv-20170807t005946-20170807t010011-006827-00c043-006.xml   S1B_OPER_AUX_POEORB_OPOD_20170827T111328_V20170806T225942_20170808T005942.EOF
	s1b-iw3-slc-vv-20170819t005946-20170819t010012-007002-00c561-006.tiff  S1B_OPER_AUX_POEORB_OPOD_20170908T111325_V20170818T225942_20170820T005942.EOF
	s1b-iw3-slc-vv-20170819t005946-20170819t010012-007002-00c561-006.xml   S1B_OPER_AUX_POEORB_OPOD_20170920T111340_V20170830T225942_20170901T005942.EOF
	s1b-iw3-slc-vv-20170831t005947-20170831t010012-007177-00ca6a-006.tiff		
	

The *tiff files are the measurements, the *xml files contain satellite information during the measurements and the *EOF files contain the orbits. Now you'll need to link the data to the orbits files, which happens in file in the format data:orbit (note the colon) in between. The first line of a file that may be called data.in in directory raw would look like this:

		s1b-iw3-slc-vv-20170515t005941-20170515t010006-005602-009d02-006:S1B_OPER_AUX_POEORB_OPOD_20170604T111305_V20170514T225942_20170516T005942.EOF
	

Note that the data files does NOT include the file extensions xml, tiff! That's important.

Now, create this file mapping all the 10 observations to their orbits. Then run (inside the raw directory):

	$> preproc_batch_tops.csh data.in ../topo/dem.grd 1
	

This creates a plot showing spatio-temporal distances between all the observations, which is called baseline.ps, look at this and start thinking about how to set up short baseline pairs. For instance, S1A20170515_ALL_F3-S1A20170714_ALL_F3 would be one pair. Note that the _F3 identifies these as Frame 3 images, so we're not processing all 3 frames in these SAR observations, but the tiff, xml files are for Frame 3 only. That's speeding things up a bit.

In the meantime, run this (takes a while!):

	$> preproc_batch_tops.csh data.in ../topo/dem.grd 2 
	

The command generates files (as in mode 1) necessary for the processing of the interferogram stack. Particularly the geometric alignment. While this is running you can start to setup the small baseline pairs. Move out of raw into the parent directory and create the file intf.in in which you add into each line the master:slave image pairs that you would like to process in a Small-baseline sense like so:

		S1A20170515_ALL_F3:S1A20170714_ALL_F3
	

Once that's done, we're almost ready to start the processing. You'll only need to get a config file and change parameters:

	$> cp $GMTSAR/gmtsar/csh/batch_tops.config ./
	

Gets you the config file. It it, set the super master image (master_image = ) to an acquisition somewhere in the center (spatially and temporally) of the baseline plot. Also, set the snaphu_threshold to 0.2, such that the interferograms will be unwrapped. Next, call intf_tops.csh with these two files:

	Usage: intf_tops.csh intf.in batch_tops.config
	  generate interferograms for a set of tops images in intf.in, dem required in ./topo
	  supermaster's name required in batch_tops.config

	  format of data.in:
	    master_image_stem:slave_image_stem

	  example of data.in
	    S1A20150628_ALL_F1:S1A20150720_ALL_F1
	    S1A20150720_ALL_F1:S1A20150809_ALL_F1

	  outputs:
	    to ./intf_all
	

Once the processing of all interferograms is done, look at all the results:

	$> eog */phase_mask_ll.png
	

These are the individual, wrapped interferograms. You can also look at the unwrapped ones (unwrap_mask_ll.png). Are there some interferograms that appear to have more signal than others? What's the location of this dataset (maybe experiment with some of the resultant kml files), what could be the cause of deformation (any guesses are good)?

This is all we'll do for this week; preparation of the time series analysis (which is time consuming). Next week you'll do the actual SBAS processing to reduce the noise and tease out the signal.

2. Getting Data

Just like with any data type: hunting is encouraged! The complexities involved with SAR data are that some data are only commercially available. Most are distributed over several archives, and the search interfaces are some times not quite intuitive. Download of the results may also not be very straightforward, but most places support command line based bulk downloads.

The algorithm that I would encourage for data discovery is:

Most places require you to register. But use is free (or commercial data are clearly marked). Some enable guest searches without registration. Register with ASF as you will need data access for homework #2. (Feel free to register with WinSAR. NMT is a Full Institutional Member. In fact, you should if you plan on becoming an active InSAR user. However, you'll have to fill out some license agreement forms and mail them to UNACVO; so this may take a few days.)

I highly recommend for you to read through the "Getting Started" guide at ASF.

Here are the links to some search interfaces:

Note that the WinSAR link takes you the "Seamless SAR Archive" (SSARA), which attempts to provide data from multiple collections (WinSAR/UNACVO, Geohazard Supersites, ASF, etc), however, the search may appear a bit rough.

For the following task keep in mind that in order to use SAR scenes for InSAR analysis, the scenes must have the same path/track (and potentially frame numbers). Using the ASF or WinSAR interfaces, search for data over an area and time of interest to you. Don't make the search area / time frame too large or you will have too many scenes to deal with.

If your search didn't bring up anything, or you are not satisfied with the number of results, you can always check with the respective space agency. To give this a try find answers to the above questions for ALOS-2 PALSAR, ALOS-PALSAR by searching the JAXA archive for your area of interest. If you prefer Sentinel-1A data, search at ESA!

Deliverables: (submit via canvas as zip archive!)

rg <at> nmt <dot> edu | Last modified: October 18 2017 19:47.