GLIMS Update: GLIMS ASTER image coverage and research documentation

[Graham Cogley]

I have been looking forward eagerly to the ASTER global DEM, which as Frank
points out is a resource of enormous potential for glaciology. But it won't
be much use to *anyone* if it relies on non-GLIMS gain settings above the
snowline. Have the global-DEM people figured this out yet? I don't
understand why there has to be a conflict between GLIMS and the DEM in this
respect.

Graham.

J. Graham Cogley, Ph.D., Professor of Geography,
Department of Geography, Trent University,
Peterborough, Ontario, CANADA K9J 7B8.

Tel    +1 705-748-1011-x7686
Fax    +1 705-742-2131
Email  gcogley at trentu.ca
Web    http://www.trentu.ca/geography/glaciology/glaciology.htm


-----Original Message-----
From: Frank Paul [mailto:fpaul at geo.uzh.ch] 
Sent: May-25-09 7:46 AM
To: gcogley1 at cogeco.ca; Jeffrey Kargel
Cc: GLIMS mailing list; stearns at ku.edu; Gordon Hamilton;
william.sneedjr at maine.edu; Slawek Tulaczyk
Subject: Re: GLIMS Update: GLIMS ASTER image coverage and research
documentation

22/tw060301
X-Qmail-Scanner-Scaninfo: 
  (Processed in 0.178242 secs by pid 20785)
X-Qmail-Scanner-Mail-From: fpaul at geo.uzh.ch via cheese
X-Qmail-Scanner-Clear: RC:1(130.60.176.65):.
Sender: owner-glims at flagmail.wr.usgs.gov
Precedence: bulk
X-Virus-Scanned: amavisd-new at trentu.ca
X-PMX-Version: 5.4.1.325704, Antispam-Engine: 2.6.0.325393, Antispam-Data:
2009.2.26.200140
X-PerlMx-Spam: Gauge=IIIIIII, Probability=7%, Report='BODY_SIZE_3000_3999 0,
BODY_SIZE_5000_LESS 0, BODY_SIZE_7000_LESS 0, INVALID_MSGID_NO_FQDN 0,
__BOUNCE_CHALLENGE_SUBJ 0, __CP_MEDIA_BODY 0, __CT 0, __CT_TEXT_PLAIN 0,
__HAS_MSGID 0, __MIME_TEXT_ONLY 0, __MIME_VERSION 0, __SANE_MSGID 0'



Dear Jeff and all

Thank you for the extended comments on GLIMS acquisitions by ASTER. I am
sympathetic to most of your points and would like to add a few general
remarks that might help to constrain the issue. From my point of view the
use of ASTER data in GLIMS can be roughly classified in three groups:

(1) Glacier outline mapping for the purpose of a global glacier inventory
(`snapshot' over a ca. 5 year period) -> must be largely cloud free & from
the end of the ablation period/dry season
(2) Scientific research projects (albedo, snow cover, lakes, debris, ...)
focusing on specific regions / individual glaciers -> could be partly
cloudy but should be taken at a specific time
(3) Repeat imagery from rapidly changing targets (e.g. glacier
lakes/hazard situations, Greenland outlet glaciers). -> needs frequent
update of many but small regions

I further identify three huge changes in the past 5 years that were not
clear during the set-up of GLIMS and ASTER STARs:
(A) freely available and already orthorectified Landsat scenes from the
USGS archive: This makes the use of ASTER data for purpose (1) nearly
obsolete
(B) rapid changes of Greenland's outlet glaciers: these regions should be
clearly specified and regionally limited for ASTER data acquisitions
(C) massive downwasting of glaciers in many parts of the world that could
be quantified from repeat DEMs: ASTER DEMs will provide an increasingly
valuable resource for this purpose

Two additional points have to be kept in mind regarding the sensor
capabilities:
ASTER has a 60 km swath with 15 m resolution while Landsat offers 180 km
with 30 m pixels (or 15 m for ETM+ pan). In this regard I would say that
ASTER is not the appropriate sensor for monitoring the two ice sheets and
their outlet glaciers or ice streams. Moreover, Landsat7 data are
available (orthorectified!) a few days after acquisition which is in my
opinion an extremely valuable resource for the monitoring of Greenlands
outlet glaciers (despite the scan line corrector stripes).

On the other hand, the proposed annual interval for glacier mapping or
even one good image in 5 years was maybe over-optimistic. While such a
frequency might work well in dry regions under constant high-pressure
systems like in Greenland, it does not work out for maritime (Alaska,
Norway, Patagonia, New Zealand), tropical or monsoon type climates. Even
in the Alps the frequency is not higher than every 5 years (with Landsat
coverage!). As far as I know, after 10 years of acquisition we have now
two strips from ASTER (23 Aug 03 & 8 Sep 04) that could be used for
purpose (1) and maybe a dozen strips for purpose (2). Please also keep in
mind that it took 22 years (Sep 2006) before a useful image for purpose
(1) of Jostedalsbreen (Norway) was acquired by Landsat 5 ...

To conclude, I am in favour of Jeff's suggestion to prepare an update of
ASTER data that has been used for GLIMS so far by all RCs (indicating
whether the purpose was 1, 2 or 3 from above) and provide information on
the regions that are still lacking appropriate coverage. For Greenlands
outlet glaciers (purpose C) I propose to identify a set of key regions
(shape file polygons) that are of utmost importance for frequent
monitoring with ASTER that are hardly covered (e.g. due to severe
striping) by Landsat 7. Please keep in X-Qmail-Scanner-Version: 1mind that
Landsat 7 ETM+ could
not be used since 2003 for purpose (1) and only partly for (2).
Hope this helps and looking forward to your feedback!

With best regards, Frank


PS: sorry for the long email :)
PPS: The upcoming ASTER GDEM will be the most important breakthrough in
creating a detailed glacier inventory (i.e. incl. topographic attributes)
from all glaciers and icecaps north of the SRTM3 coverage.