17 April 2016

2016 Tough on Sea Ice Satellites

The last several weeks have been hard on the satellites people like me use most for determining sea ice coverage.  We use passive microwave instruments on a number of different satellites.  The 'passive' in its name means that it doesn't emit microwaves.  It just sits back and collects the emissions from where it's looking.  In this, the instrument is rather like our eyes.

Different centers use different instruments and different combinations of instruments.  The main ones are:
SSMI on DMSP F-15, launched on 12 December 1999 (pushing 17 years!)
SSMI-S on DMSP F-16, launched 18 October 2003 (pushing 13!)
SSMI-S on DMSP F-17, 4 November 2006 (almost 10!)
SSMI-S on DMSP F-18, 18 October 2009 (approaching 7)
SSMI-S on DMSP F-19, 3 April 2014 (only 2)
AMSR2 on GCOM-W, 18 May 2012 (nearly 4)

A word about the names.  SSMI is Special Sensor Microwave Imager.  'Imager' is the key word.  With satellites, 'imager' means that the instrument is designed to be able to see (mostly) the surface.  Handy for us sea ice people.  DMSP is the Defense Meteorological Satellite Program -- the US Department of Defense operates these satellites.  SSMI-S (or SSMI-SU) is the SSMI -- Sounder (or Sounding Unit).  Means that in addition to the regular SSMI observing of the surface, it also carries some sensors that can do 'sounding'.  Sounding is to see what's going on in the atmosphere rather than mostly the surface.  (Name comes from the weather balloons -- which collect data known as soundings.)  AMSR is Advanced Microwave Sounding Radiometer.  It's operated by the Japanese Space Agency (JAXA).  The advance is that it is able to see more detail and the much older designs in the SSMI and SSMI-S. 

So, to our stories of woe.  All of these instruments are designed for 5 years' operation.  F-15 giving (mostly) good data after 17 years is spectacular for this type of satellite.  Notice that most of these potential data sources are already past their design life.  Since February 2016:

F-16: the sounding channels quit working early February
F-17: April 5th data quality impaired on one of the surface imaging channels, data volumes sent are greatly reduced.
F-18: Mostly ok, but reduced volume of data.  Many orbits' data not making it through.
F-19: Data ceased flowing February 2
AMSR2: Data outage afternoon of April 15th through morning of April 16th.

F-15, the oldest of the crowd, is still sending basically normal data volumes at basically normal volumes.

So, hiccups all around the sea ice analysis world.  The NSIDC was using only the F-17 SSMI-S, so has to rebuild their system to work with another instrument.  The AMSR2 temporary outage affected some centers seriously as they relied only on that instrument.  The US NWS uses both F-15 and F-17, and so far seems to be ok.  I haven't checked the operating status of the OSI-SAF sea ice (European analysis).  If I remember correctly, they also use more than one instrument, so should also be ok.

More gory details below the fold ...
Even the instruments which are providing mostly good data are imperfect.  F-15's data at 22 GHz is corrupted.  Most sea ice analysis systems use this for weather filtering.  (I constructed a different weather filter to avoid this bad channel at my workplace.)  F-16 hasn't been as widely used in the sea ice world as 15 or 17, so losing it makes for less difficulty immediately, but does mean that we have fewer options to move away from 15 or 17.  F-17 is/was widely used, so many places are having difficulty.  F-18 has a very bad channel at 150 GHz, but the other imaging channels seem to be ok.  F-19 should be the best of the group, being the youngest, but quit providing data. 

Aside from the April 15-16 outage, AMSR2 is or should still be the best -- still within its design life, and with a more advanced design.  A lot of people in the sea ice world are probably hoping strongly that this outage is a temporary thing, not to be repeated.

The challenge about using a different instrument is that they're not identical.  If you're trying to study climate, as many centers are, small differences in the details of what the sensor sees can make for large changes in the trends you'd fine in a sea ice analysis.  So you have to work, hard, to make sure that the analysis you get from AMSR2, for instance, is consistent with what you were getting from F-17 for the last several years.  This is what the NSIDC is likely doing right now.  Since my professional concern is more weather, I don't have quite the issue.  But I still need data to work with, and losing perhaps 2 potential backups is not cheery.  Losing half the data from one of the satellites I already use is also hot happy.  And, although my quality control procedures (or those which are cutting my supply in half) are doing ok for now, it's uncomfortable to know that my data supply is even less good than it used to be.

It's for reasons like this that all the people I know who work with data agree immediately with the comment "All data have problems."


Hank Roberts said...


Robert Grumbine said...

Yes, a continuing issue. Back in the 90's (also) there was a representative in congress of the same opinion -- we didn't need the National Weather Service because he got his weather from the Weather Channel. This is a bit ironic, given that the Weather Channel is quite up front about the fact that they get their information from NOAA / National Weather Service.

Jim Hunt said...

Do you suppose there is now a possibility that F-20 will be un-mothballed?

Robert Grumbine said...

Possibility, sure. But it's a decision in Congress (which ordered the mothballing) rather than the Defense Department, so make your own election year prediction as to probability and timing.

Jim Hunt said...

I'm from the UK Bob, and hence not entirely au fait with how to go about making US presidential election predictions! However I did read this a while ago:


Would you agree that "The Air Force has not properly managed its space weather program"?

EliRabett said...

Satellites and space probes are like people, infant mortality is a problem (as in crash and burn or never wake up) they perform so well that they often exceed their design lifetime. Indeed, this is often a problem as a working satellite is expensive to operate and the choice becomes to turn one off and launch a new one or to keep on operating the old one.

NASA has on occasion offered operations of some satellites for free to any organization that really cares. The Mars Rovers are examples of long operating probes that have a constituency which does not allow them to be turned off.